{"cells":[{"cell_type":"markdown","source":["Reference:\n","Raschka et al.\n","\n","https://cfteach.github.io/NNDL_DATA621/referencesmc.html"],"metadata":{"id":"rnDaJ8L5iYbU"}},{"cell_type":"markdown","metadata":{"tags":[],"id":"zi3m10hGiEz5"},"source":["## Code examples for Data Preprocessing"]},{"cell_type":"code","source":["import numpy as np"],"metadata":{"id":"u3hGqEIAizR1","executionInfo":{"status":"ok","timestamp":1726116217530,"user_tz":240,"elapsed":175,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}}},"execution_count":6,"outputs":[]},{"cell_type":"code","execution_count":1,"metadata":{"id":"8i-yYHvfiEz_","executionInfo":{"status":"ok","timestamp":1726116164329,"user_tz":240,"elapsed":182,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}}},"outputs":[],"source":["from IPython.display import Image\n","%matplotlib inline"]},{"cell_type":"markdown","metadata":{"id":"3gtHRiwFiE0A"},"source":["### Dealing with missing data"]},{"cell_type":"markdown","metadata":{"id":"Fw5fNBxDiE0A"},"source":["Identifying missing values in tabular data"]},{"cell_type":"code","execution_count":2,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":160},"id":"CiD6yAv2iE0A","executionInfo":{"status":"ok","timestamp":1726116200897,"user_tz":240,"elapsed":1051,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"b21d7b4e-1c4f-40fd-db68-0ccbb25e9a12"},"outputs":[{"output_type":"execute_result","data":{"text/plain":[" A B C D\n","0 1.0 2.0 3.0 4.0\n","1 5.0 6.0 NaN 8.0\n","2 10.0 11.0 12.0 NaN"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
ABCD
01.02.03.04.0
15.06.0NaN8.0
210.011.012.0NaN
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n"," \n"," \n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"df","summary":"{\n \"name\": \"df\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"A\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 4.509249752822894,\n \"min\": 1.0,\n \"max\": 10.0,\n \"num_unique_values\": 3,\n \"samples\": [\n 1.0,\n 5.0,\n 10.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"B\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 4.509249752822894,\n \"min\": 2.0,\n \"max\": 11.0,\n \"num_unique_values\": 3,\n \"samples\": [\n 2.0,\n 6.0,\n 11.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"C\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 6.363961030678928,\n \"min\": 3.0,\n \"max\": 12.0,\n \"num_unique_values\": 2,\n \"samples\": [\n 12.0,\n 3.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"D\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 2.8284271247461903,\n \"min\": 4.0,\n \"max\": 8.0,\n \"num_unique_values\": 2,\n \"samples\": [\n 8.0,\n 4.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":2}],"source":["import pandas as pd\n","from io import StringIO\n","import sys\n","\n","csv_data = \\\n","'''A,B,C,D\n","1.0,2.0,3.0,4.0\n","5.0,6.0,,8.0\n","10.0,11.0,12.0,'''\n","\n","# If you are using Python 2.7, you need\n","# to convert the string to unicode:\n","\n","if (sys.version_info < (3, 0)):\n"," csv_data = unicode(csv_data)\n","\n","# StringIO simulates a file=like object in memory,\n","#like if it was a regular CSV file to read from the hard drive\n","df = pd.read_csv(StringIO(csv_data))\n","\n","df"]},{"cell_type":"code","source":["df.isnull()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":160},"id":"aUb4F5rikqg_","executionInfo":{"status":"ok","timestamp":1726116204221,"user_tz":240,"elapsed":207,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"a7e54ce6-0dfc-4bf4-eb80-501333db4dd3"},"execution_count":3,"outputs":[{"output_type":"execute_result","data":{"text/plain":[" A B C D\n","0 False False False False\n","1 False False True False\n","2 False False False True"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
ABCD
0FalseFalseFalseFalse
1FalseFalseTrueFalse
2FalseFalseFalseTrue
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","summary":"{\n \"name\": \"df\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"A\",\n \"properties\": {\n \"dtype\": \"boolean\",\n \"num_unique_values\": 1,\n \"samples\": [\n false\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"B\",\n \"properties\": {\n \"dtype\": \"boolean\",\n \"num_unique_values\": 1,\n \"samples\": [\n false\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"C\",\n \"properties\": {\n \"dtype\": \"boolean\",\n \"num_unique_values\": 2,\n \"samples\": [\n true\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"D\",\n \"properties\": {\n \"dtype\": \"boolean\",\n \"num_unique_values\": 2,\n \"samples\": [\n true\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":3}]},{"cell_type":"code","execution_count":4,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":224},"id":"FG0n701NiE0C","executionInfo":{"status":"ok","timestamp":1726116205049,"user_tz":240,"elapsed":252,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"21768c61-bbde-492f-9ea8-b6ae2287c46a"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["A 0\n","B 0\n","C 1\n","D 1\n","dtype: int64"],"text/html":["
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
0
A0
B0
C1
D1
\n","

"]},"metadata":{},"execution_count":4}],"source":["df.isnull().sum(axis=0)"]},{"cell_type":"code","source":["data = pd.Series([1, np.nan, 3, None, 5])\n","print(data.isnull())"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"VWe6DoXhc38g","executionInfo":{"status":"ok","timestamp":1726116221701,"user_tz":240,"elapsed":168,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"a55b2b5f-d89c-4b37-93ae-1ce1062ba8ef"},"execution_count":7,"outputs":[{"output_type":"stream","name":"stdout","text":["0 False\n","1 True\n","2 False\n","3 True\n","4 False\n","dtype: bool\n"]}]},{"cell_type":"code","execution_count":8,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"gPG_CzXiiE0C","executionInfo":{"status":"ok","timestamp":1726116227188,"user_tz":240,"elapsed":167,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"f72a0723-1624-4803-f9a9-3da4e483bd7b"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[ 1., 2., 3., 4.],\n"," [ 5., 6., nan, 8.],\n"," [10., 11., 12., nan]])"]},"metadata":{},"execution_count":8}],"source":["# access the underlying NumPy array\n","# via the `values` attribute\n","df.values"]},{"cell_type":"code","source":["data = np.array([1, np.nan, 3, None], dtype=float)\n","data\n","\n","print(np.isnan(data))"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"G-tLaisPfYRY","executionInfo":{"status":"ok","timestamp":1726116257798,"user_tz":240,"elapsed":185,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"f772663e-1e0e-4a19-b15f-9556389e3d89"},"execution_count":9,"outputs":[{"output_type":"stream","name":"stdout","text":["[False True False True]\n"]}]},{"cell_type":"markdown","metadata":{"id":"Z2zDbPOIiE0C"},"source":["
\n","
"]},{"cell_type":"markdown","metadata":{"id":"jENCqu4eiE0D"},"source":["Eliminating training examples or features with missing values"]},{"cell_type":"code","source":["df"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":160},"id":"ilUD23mBp00a","executionInfo":{"status":"ok","timestamp":1726116270713,"user_tz":240,"elapsed":198,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"cbb75cc6-ce9d-4862-a1a0-bd3294f26a1d"},"execution_count":10,"outputs":[{"output_type":"execute_result","data":{"text/plain":[" A B C D\n","0 1.0 2.0 3.0 4.0\n","1 5.0 6.0 NaN 8.0\n","2 10.0 11.0 12.0 NaN"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
ABCD
01.02.03.04.0
15.06.0NaN8.0
210.011.012.0NaN
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n"," \n"," \n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"df","summary":"{\n \"name\": \"df\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"A\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 4.509249752822894,\n \"min\": 1.0,\n \"max\": 10.0,\n \"num_unique_values\": 3,\n \"samples\": [\n 1.0,\n 5.0,\n 10.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"B\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 4.509249752822894,\n \"min\": 2.0,\n \"max\": 11.0,\n \"num_unique_values\": 3,\n \"samples\": [\n 2.0,\n 6.0,\n 11.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"C\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 6.363961030678928,\n \"min\": 3.0,\n \"max\": 12.0,\n \"num_unique_values\": 2,\n \"samples\": [\n 12.0,\n 3.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"D\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 2.8284271247461903,\n \"min\": 4.0,\n \"max\": 8.0,\n \"num_unique_values\": 2,\n \"samples\": [\n 8.0,\n 4.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":10}]},{"cell_type":"code","execution_count":11,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":100},"id":"jC3kt_qOiE0D","executionInfo":{"status":"ok","timestamp":1726116273325,"user_tz":240,"elapsed":161,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"ef53cfc9-5bdc-4c0b-c4ba-b3ac4eb444e9"},"outputs":[{"output_type":"execute_result","data":{"text/plain":[" A B C D\n","0 1.0 2.0 3.0 4.0"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
ABCD
01.02.03.04.0
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","summary":"{\n \"name\": \"df\",\n \"rows\": 1,\n \"fields\": [\n {\n \"column\": \"A\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": null,\n \"min\": 1.0,\n \"max\": 1.0,\n \"num_unique_values\": 1,\n \"samples\": [\n 1.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"B\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": null,\n \"min\": 2.0,\n \"max\": 2.0,\n \"num_unique_values\": 1,\n \"samples\": [\n 2.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"C\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": null,\n \"min\": 3.0,\n \"max\": 3.0,\n \"num_unique_values\": 1,\n \"samples\": [\n 3.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"D\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": null,\n \"min\": 4.0,\n \"max\": 4.0,\n \"num_unique_values\": 1,\n \"samples\": [\n 4.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":11}],"source":["# remove rows that contain missing values\n","\n","df.dropna(axis=0)"]},{"cell_type":"code","execution_count":12,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":160},"id":"oZpnL29LiE0D","executionInfo":{"status":"ok","timestamp":1726116275137,"user_tz":240,"elapsed":175,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"e1a1c391-7446-42d1-c824-507ed78d85e6"},"outputs":[{"output_type":"execute_result","data":{"text/plain":[" A B\n","0 1.0 2.0\n","1 5.0 6.0\n","2 10.0 11.0"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
AB
01.02.0
15.06.0
210.011.0
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","summary":"{\n \"name\": \"df\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"A\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 4.509249752822894,\n \"min\": 1.0,\n \"max\": 10.0,\n \"num_unique_values\": 3,\n \"samples\": [\n 1.0,\n 5.0,\n 10.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"B\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 4.509249752822894,\n \"min\": 2.0,\n \"max\": 11.0,\n \"num_unique_values\": 3,\n \"samples\": [\n 2.0,\n 6.0,\n 11.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":12}],"source":["# remove columns that contain missing values\n","\n","df.dropna(axis=1)"]},{"cell_type":"code","execution_count":13,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":160},"id":"EgB7f349iE0E","executionInfo":{"status":"ok","timestamp":1726116279741,"user_tz":240,"elapsed":203,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"2acd88b7-25cd-468a-cf5c-6c5391cfed35"},"outputs":[{"output_type":"execute_result","data":{"text/plain":[" A B C D\n","0 1.0 2.0 3.0 4.0\n","1 5.0 6.0 NaN 8.0\n","2 10.0 11.0 12.0 NaN"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
ABCD
01.02.03.04.0
15.06.0NaN8.0
210.011.012.0NaN
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","summary":"{\n \"name\": \"df\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"A\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 4.509249752822894,\n \"min\": 1.0,\n \"max\": 10.0,\n \"num_unique_values\": 3,\n \"samples\": [\n 1.0,\n 5.0,\n 10.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"B\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 4.509249752822894,\n \"min\": 2.0,\n \"max\": 11.0,\n \"num_unique_values\": 3,\n \"samples\": [\n 2.0,\n 6.0,\n 11.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"C\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 6.363961030678928,\n \"min\": 3.0,\n \"max\": 12.0,\n \"num_unique_values\": 2,\n \"samples\": [\n 12.0,\n 3.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"D\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 2.8284271247461903,\n \"min\": 4.0,\n \"max\": 8.0,\n \"num_unique_values\": 2,\n \"samples\": [\n 8.0,\n 4.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":13}],"source":["# only drop rows where all columns are NaN\n","\n","df.dropna(how='all')"]},{"cell_type":"code","execution_count":15,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":100},"id":"qNSUZjgTiE0F","executionInfo":{"status":"ok","timestamp":1726116313049,"user_tz":240,"elapsed":218,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"b3faa7c5-3c3c-40ae-eacd-46f70acffaf0"},"outputs":[{"output_type":"execute_result","data":{"text/plain":[" A B C D\n","0 1.0 2.0 3.0 4.0"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
ABCD
01.02.03.04.0
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","summary":"{\n \"name\": \"df\",\n \"rows\": 1,\n \"fields\": [\n {\n \"column\": \"A\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": null,\n \"min\": 1.0,\n \"max\": 1.0,\n \"num_unique_values\": 1,\n \"samples\": [\n 1.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"B\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": null,\n \"min\": 2.0,\n \"max\": 2.0,\n \"num_unique_values\": 1,\n \"samples\": [\n 2.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"C\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": null,\n \"min\": 3.0,\n \"max\": 3.0,\n \"num_unique_values\": 1,\n \"samples\": [\n 3.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"D\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": null,\n \"min\": 4.0,\n \"max\": 4.0,\n \"num_unique_values\": 1,\n \"samples\": [\n 4.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":15}],"source":["# drop rows that have fewer than 4 real values\n","\n","df.dropna(thresh=4)"]},{"cell_type":"code","execution_count":16,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":130},"id":"fHmJjUr4iE0F","executionInfo":{"status":"ok","timestamp":1726116323933,"user_tz":240,"elapsed":173,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"63c05fff-4b54-4d0b-c834-7841208a2c53"},"outputs":[{"output_type":"execute_result","data":{"text/plain":[" A B C D\n","0 1.0 2.0 3.0 4.0\n","2 10.0 11.0 12.0 NaN"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
ABCD
01.02.03.04.0
210.011.012.0NaN
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","summary":"{\n \"name\": \"df\",\n \"rows\": 2,\n \"fields\": [\n {\n \"column\": \"A\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 6.363961030678928,\n \"min\": 1.0,\n \"max\": 10.0,\n \"num_unique_values\": 2,\n \"samples\": [\n 10.0,\n 1.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"B\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 6.363961030678928,\n \"min\": 2.0,\n \"max\": 11.0,\n \"num_unique_values\": 2,\n \"samples\": [\n 11.0,\n 2.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"C\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 6.363961030678928,\n \"min\": 3.0,\n \"max\": 12.0,\n \"num_unique_values\": 2,\n \"samples\": [\n 12.0,\n 3.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"D\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": null,\n \"min\": 4.0,\n \"max\": 4.0,\n \"num_unique_values\": 1,\n \"samples\": [\n 4.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":16}],"source":["# only drop rows where NaN appear in specific columns (here: 'C')\n","\n","df.dropna(subset=['C'])"]},{"cell_type":"markdown","metadata":{"id":"tFFjCLeYiE0F"},"source":["
\n","
"]},{"cell_type":"markdown","metadata":{"id":"8R8dg_M2iE0F"},"source":["### Imputing missing values"]},{"cell_type":"code","execution_count":17,"metadata":{"scrolled":true,"colab":{"base_uri":"https://localhost:8080/"},"id":"LK1ACtltiE0F","executionInfo":{"status":"ok","timestamp":1726116370355,"user_tz":240,"elapsed":312,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"ee43d5d7-aa8b-4976-a2f3-f7e6db1430da"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[ 1., 2., 3., 4.],\n"," [ 5., 6., nan, 8.],\n"," [10., 11., 12., nan]])"]},"metadata":{},"execution_count":17}],"source":["# again: our original array\n","df.values"]},{"cell_type":"code","execution_count":18,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"eZ9X3V6ViE0G","executionInfo":{"status":"ok","timestamp":1726116374902,"user_tz":240,"elapsed":2764,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"06dfe212-d444-4ced-949e-16350552c04e"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[ 1. , 2. , 3. , 4. ],\n"," [ 5. , 6. , 7.5, 8. ],\n"," [10. , 11. , 12. , 6. ]])"]},"metadata":{},"execution_count":18}],"source":["# impute missing values via the column mean\n","\n","from sklearn.impute import SimpleImputer\n","import numpy as np\n","\n","imr = SimpleImputer(missing_values=np.nan, strategy='mean') #other methods: 'median', 'most_frequent'\n","imr = imr.fit(df.values)\n","imputed_data = imr.transform(df.values)\n","imputed_data"]},{"cell_type":"markdown","metadata":{"id":"kvqqZY50iE0G"},"source":["
\n","
"]},{"cell_type":"code","execution_count":19,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":160},"id":"pwk0ZUq1iE0G","executionInfo":{"status":"ok","timestamp":1726116377869,"user_tz":240,"elapsed":174,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"8000549b-6ab0-4bae-f17b-76c620332072"},"outputs":[{"output_type":"execute_result","data":{"text/plain":[" A B C D\n","0 1.0 2.0 3.0 4.0\n","1 5.0 6.0 7.5 8.0\n","2 10.0 11.0 12.0 6.0"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
ABCD
01.02.03.04.0
15.06.07.58.0
210.011.012.06.0
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","summary":"{\n \"name\": \"df\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"A\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 4.509249752822894,\n \"min\": 1.0,\n \"max\": 10.0,\n \"num_unique_values\": 3,\n \"samples\": [\n 1.0,\n 5.0,\n 10.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"B\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 4.509249752822894,\n \"min\": 2.0,\n \"max\": 11.0,\n \"num_unique_values\": 3,\n \"samples\": [\n 2.0,\n 6.0,\n 11.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"C\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 4.5,\n \"min\": 3.0,\n \"max\": 12.0,\n \"num_unique_values\": 3,\n \"samples\": [\n 3.0,\n 7.5,\n 12.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"D\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 2.0,\n \"min\": 4.0,\n \"max\": 8.0,\n \"num_unique_values\": 3,\n \"samples\": [\n 4.0,\n 8.0,\n 6.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":19}],"source":["df.fillna(df.mean())"]},{"cell_type":"code","source":["df"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":160},"id":"9UP87hDukSf8","executionInfo":{"status":"ok","timestamp":1726116606708,"user_tz":240,"elapsed":185,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"02633944-75e0-4fa4-aae8-4b5a62a9865d"},"execution_count":24,"outputs":[{"output_type":"execute_result","data":{"text/plain":[" A B C D\n","0 1.0 2.0 3.0 4.0\n","1 5.0 6.0 NaN 8.0\n","2 10.0 11.0 12.0 NaN"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
ABCD
01.02.03.04.0
15.06.0NaN8.0
210.011.012.0NaN
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n"," \n"," \n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"df","summary":"{\n \"name\": \"df\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"A\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 4.509249752822894,\n \"min\": 1.0,\n \"max\": 10.0,\n \"num_unique_values\": 3,\n \"samples\": [\n 1.0,\n 5.0,\n 10.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"B\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 4.509249752822894,\n \"min\": 2.0,\n \"max\": 11.0,\n \"num_unique_values\": 3,\n \"samples\": [\n 2.0,\n 6.0,\n 11.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"C\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 6.363961030678928,\n \"min\": 3.0,\n \"max\": 12.0,\n \"num_unique_values\": 2,\n \"samples\": [\n 12.0,\n 3.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"D\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 2.8284271247461903,\n \"min\": 4.0,\n \"max\": 8.0,\n \"num_unique_values\": 2,\n \"samples\": [\n 8.0,\n 4.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":24}]},{"cell_type":"code","source":["dg = df.copy()"],"metadata":{"id":"d50P1ou1ns2l","executionInfo":{"status":"ok","timestamp":1726116386450,"user_tz":240,"elapsed":203,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}}},"execution_count":20,"outputs":[]},{"cell_type":"code","source":["row, col = (dg == 5.0).stack().idxmax()\n","print(row, col)"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"QquvBoianv7o","executionInfo":{"status":"ok","timestamp":1726116387318,"user_tz":240,"elapsed":155,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"12c9b67e-692a-4190-cb25-b693df76a0c2"},"execution_count":21,"outputs":[{"output_type":"stream","name":"stdout","text":["1 A\n"]}]},{"cell_type":"code","source":["dg.loc[0,'A'] = None"],"metadata":{"id":"VqrjruRYosVc","executionInfo":{"status":"ok","timestamp":1726116393783,"user_tz":240,"elapsed":188,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}}},"execution_count":22,"outputs":[]},{"cell_type":"code","source":["dg"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":160},"id":"r_c-luCbo0WW","executionInfo":{"status":"ok","timestamp":1726116400071,"user_tz":240,"elapsed":169,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"383052cf-bf70-4b28-f530-eac34e4f415f"},"execution_count":23,"outputs":[{"output_type":"execute_result","data":{"text/plain":[" A B C D\n","0 NaN 2.0 3.0 4.0\n","1 5.0 6.0 NaN 8.0\n","2 10.0 11.0 12.0 NaN"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
ABCD
0NaN2.03.04.0
15.06.0NaN8.0
210.011.012.0NaN
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n"," \n"," \n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"dg","summary":"{\n \"name\": \"dg\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"A\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 3.5355339059327378,\n \"min\": 5.0,\n \"max\": 10.0,\n \"num_unique_values\": 2,\n \"samples\": [\n 10.0,\n 5.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"B\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 4.509249752822894,\n \"min\": 2.0,\n \"max\": 11.0,\n \"num_unique_values\": 3,\n \"samples\": [\n 2.0,\n 6.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"C\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 6.363961030678928,\n \"min\": 3.0,\n \"max\": 12.0,\n \"num_unique_values\": 2,\n \"samples\": [\n 12.0,\n 3.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"D\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 2.8284271247461903,\n \"min\": 4.0,\n \"max\": 8.0,\n \"num_unique_values\": 2,\n \"samples\": [\n 8.0,\n 4.0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":23}]},{"cell_type":"markdown","metadata":{"id":"tUV5VyOqiE0H"},"source":["
\n","
"]},{"cell_type":"markdown","metadata":{"id":"yCcp7bNuiE0H"},"source":["### Handling categorical data"]},{"cell_type":"markdown","metadata":{"id":"oSG18R88iE0J"},"source":["----- Nominal and ordinal features"]},{"cell_type":"code","execution_count":25,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":160},"id":"Uh1F6FMViE0K","executionInfo":{"status":"ok","timestamp":1726116639666,"user_tz":240,"elapsed":194,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"2477db0f-997b-4475-8939-6f4ccd9de929"},"outputs":[{"output_type":"execute_result","data":{"text/plain":[" color size price classlabel\n","0 green M 10.1 class2\n","1 red L 13.5 class1\n","2 blue XL 15.3 class2"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
colorsizepriceclasslabel
0greenM10.1class2
1redL13.5class1
2blueXL15.3class2
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n"," \n"," \n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"df","summary":"{\n \"name\": \"df\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"color\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 3,\n \"samples\": [\n \"green\",\n \"red\",\n \"blue\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"size\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 3,\n \"samples\": [\n \"M\",\n \"L\",\n \"XL\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"price\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 2.640706976045115,\n \"min\": 10.1,\n \"max\": 15.3,\n \"num_unique_values\": 3,\n \"samples\": [\n 10.1,\n 13.5,\n 15.3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"classlabel\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 2,\n \"samples\": [\n \"class1\",\n \"class2\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":25}],"source":["import pandas as pd\n","\n","df = pd.DataFrame([['green', 'M', 10.1, 'class2'],\n"," ['red', 'L', 13.5, 'class1'],\n"," ['blue', 'XL', 15.3, 'class2']])\n","\n","df.columns = ['color', 'size', 'price', 'classlabel']\n","df"]},{"cell_type":"code","source":["dg = pd.DataFrame([['green','XL','class1'],['yellow','M','class1'],['red','L','class2'],['blue','S','class2']])\n","dg.columns = ['colors','size','label']\n","dg"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":191},"id":"6v3UahT7mwMX","executionInfo":{"status":"ok","timestamp":1726116642774,"user_tz":240,"elapsed":4,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"10285105-d4e8-48cd-dcc2-29e7972839e5"},"execution_count":26,"outputs":[{"output_type":"execute_result","data":{"text/plain":[" colors size label\n","0 green XL class1\n","1 yellow M class1\n","2 red L class2\n","3 blue S class2"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
colorssizelabel
0greenXLclass1
1yellowMclass1
2redLclass2
3blueSclass2
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n"," \n"," \n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"dg","summary":"{\n \"name\": \"dg\",\n \"rows\": 4,\n \"fields\": [\n {\n \"column\": \"colors\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 4,\n \"samples\": [\n \"yellow\",\n \"blue\",\n \"green\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"size\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 4,\n \"samples\": [\n \"M\",\n \"S\",\n \"XL\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"label\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 2,\n \"samples\": [\n \"class2\",\n \"class1\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":26}]},{"cell_type":"code","source":["mapping = {'XL':4,'L':3,'M':2,'S':1}"],"metadata":{"id":"g4wxVt4lnK3b","executionInfo":{"status":"ok","timestamp":1726116645472,"user_tz":240,"elapsed":170,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}}},"execution_count":27,"outputs":[]},{"cell_type":"code","source":["dg['size'] = dg['size'].map(mapping)"],"metadata":{"id":"XjhWzgAsnbZ3","executionInfo":{"status":"ok","timestamp":1726116646406,"user_tz":240,"elapsed":292,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}}},"execution_count":28,"outputs":[]},{"cell_type":"code","source":["dg"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":191},"id":"bKPtx2zpobxC","executionInfo":{"status":"ok","timestamp":1726116647521,"user_tz":240,"elapsed":202,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"c38f3101-4806-414b-b957-d88cb56fa7b7"},"execution_count":29,"outputs":[{"output_type":"execute_result","data":{"text/plain":[" colors size label\n","0 green 4 class1\n","1 yellow 2 class1\n","2 red 3 class2\n","3 blue 1 class2"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
colorssizelabel
0green4class1
1yellow2class1
2red3class2
3blue1class2
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n"," \n"," \n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"dg","summary":"{\n \"name\": \"dg\",\n \"rows\": 4,\n \"fields\": [\n {\n \"column\": \"colors\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 4,\n \"samples\": [\n \"yellow\",\n \"blue\",\n \"green\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"size\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 1,\n \"min\": 1,\n \"max\": 4,\n \"num_unique_values\": 4,\n \"samples\": [\n 2,\n 1,\n 4\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"label\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 2,\n \"samples\": [\n \"class2\",\n \"class1\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":29}]},{"cell_type":"code","source":["for k,s in mapping.items():\n"," print(k,s)"],"metadata":{"id":"yC5R-shEn3Vb","executionInfo":{"status":"ok","timestamp":1726116652984,"user_tz":240,"elapsed":181,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/"},"outputId":"788abcb4-6e0c-43cc-cfd3-0506681d3a30"},"execution_count":30,"outputs":[{"output_type":"stream","name":"stdout","text":["XL 4\n","L 3\n","M 2\n","S 1\n"]}]},{"cell_type":"code","source":["inverse_mapping = {k:s for s,k in mapping.items()}\n","inverse_mapping"],"metadata":{"id":"kCeC_9lgnmSx","executionInfo":{"status":"ok","timestamp":1726116655315,"user_tz":240,"elapsed":180,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/"},"outputId":"af3504df-b123-4a00-ff98-94926f4d3ce9"},"execution_count":31,"outputs":[{"output_type":"execute_result","data":{"text/plain":["{4: 'XL', 3: 'L', 2: 'M', 1: 'S'}"]},"metadata":{},"execution_count":31}]},{"cell_type":"code","source":["dg['size'] = dg['size'].map(inverse_mapping)"],"metadata":{"id":"vB6BjZbHoX0-","executionInfo":{"status":"ok","timestamp":1726116656763,"user_tz":240,"elapsed":223,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}}},"execution_count":32,"outputs":[]},{"cell_type":"code","source":["dg"],"metadata":{"id":"bUHXvyG6oZkP","executionInfo":{"status":"ok","timestamp":1726116657350,"user_tz":240,"elapsed":4,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/","height":191},"outputId":"78ce9780-f518-4df2-e0bf-5f7b7b977ea9"},"execution_count":33,"outputs":[{"output_type":"execute_result","data":{"text/plain":[" colors size label\n","0 green XL class1\n","1 yellow M class1\n","2 red L class2\n","3 blue S class2"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
colorssizelabel
0greenXLclass1
1yellowMclass1
2redLclass2
3blueSclass2
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n"," \n"," \n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"dg","summary":"{\n \"name\": \"dg\",\n \"rows\": 4,\n \"fields\": [\n {\n \"column\": \"colors\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 4,\n \"samples\": [\n \"yellow\",\n \"blue\",\n \"green\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"size\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 4,\n \"samples\": [\n \"M\",\n \"S\",\n \"XL\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"label\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 2,\n \"samples\": [\n \"class2\",\n \"class1\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":33}]},{"cell_type":"markdown","metadata":{"id":"Q68Iz4tyiE0L"},"source":["
\n","
"]},{"cell_type":"markdown","metadata":{"id":"Het_5_fGiE0L"},"source":["Mapping ordinal features"]},{"cell_type":"code","execution_count":34,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":160},"id":"ITfBi6V5iE0M","executionInfo":{"status":"ok","timestamp":1726116669537,"user_tz":240,"elapsed":211,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"cc8897fc-f931-4f2b-ce95-7e86460a610e"},"outputs":[{"output_type":"execute_result","data":{"text/plain":[" color size price classlabel\n","0 green 1 10.1 class2\n","1 red 2 13.5 class1\n","2 blue 3 15.3 class2"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
colorsizepriceclasslabel
0green110.1class2
1red213.5class1
2blue315.3class2
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n"," \n"," \n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"df","summary":"{\n \"name\": \"df\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"color\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 3,\n \"samples\": [\n \"green\",\n \"red\",\n \"blue\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"size\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 1,\n \"min\": 1,\n \"max\": 3,\n \"num_unique_values\": 3,\n \"samples\": [\n 1,\n 2,\n 3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"price\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 2.640706976045115,\n \"min\": 10.1,\n \"max\": 15.3,\n \"num_unique_values\": 3,\n \"samples\": [\n 10.1,\n 13.5,\n 15.3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"classlabel\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 2,\n \"samples\": [\n \"class1\",\n \"class2\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":34}],"source":["size_mapping = {'XL': 3,\n"," 'L': 2,\n"," 'M': 1}\n","\n","df['size'] = df['size'].map(size_mapping)\n","df"]},{"cell_type":"code","execution_count":35,"metadata":{"id":"RN02eeYkiE0M","executionInfo":{"status":"ok","timestamp":1726116672281,"user_tz":240,"elapsed":186,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/","height":193},"outputId":"a53331a7-12c3-40d1-8eb4-9c3645ab3c11"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["0 M\n","1 L\n","2 XL\n","Name: size, dtype: object"],"text/html":["
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
size
0M
1L
2XL
\n","

"]},"metadata":{},"execution_count":35}],"source":["inv_size_mapping = {v: k for k, v in size_mapping.items()}\n","df['size'].map(inv_size_mapping)"]},{"cell_type":"markdown","metadata":{"id":"COdHXXT9iE0M"},"source":["
\n","
"]},{"cell_type":"markdown","metadata":{"id":"6N7SoVoZiE0M"},"source":["### Encoding class labels"]},{"cell_type":"code","execution_count":36,"metadata":{"id":"IH3x29fpiE0M","executionInfo":{"status":"ok","timestamp":1726116702055,"user_tz":240,"elapsed":173,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/"},"outputId":"7cfef634-ac84-473c-d47a-b366f2817210"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["{'class1': 0, 'class2': 1}"]},"metadata":{},"execution_count":36}],"source":["import numpy as np\n","\n","# create a mapping dict\n","# to convert class labels from strings to integers\n","class_mapping = {label: idx for idx, label in enumerate(np.unique(df['classlabel']))}\n","class_mapping"]},{"cell_type":"code","execution_count":37,"metadata":{"id":"6ND7uBUDiE0N","executionInfo":{"status":"ok","timestamp":1726116703695,"user_tz":240,"elapsed":168,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/","height":160},"outputId":"9ce739a1-9928-4447-f57b-6ad4f3635644"},"outputs":[{"output_type":"execute_result","data":{"text/plain":[" color size price classlabel\n","0 green 1 10.1 1\n","1 red 2 13.5 0\n","2 blue 3 15.3 1"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
colorsizepriceclasslabel
0green110.11
1red213.50
2blue315.31
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n"," \n"," \n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"df","summary":"{\n \"name\": \"df\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"color\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 3,\n \"samples\": [\n \"green\",\n \"red\",\n \"blue\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"size\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 1,\n \"min\": 1,\n \"max\": 3,\n \"num_unique_values\": 3,\n \"samples\": [\n 1,\n 2,\n 3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"price\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 2.640706976045115,\n \"min\": 10.1,\n \"max\": 15.3,\n \"num_unique_values\": 3,\n \"samples\": [\n 10.1,\n 13.5,\n 15.3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"classlabel\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 0,\n \"min\": 0,\n \"max\": 1,\n \"num_unique_values\": 2,\n \"samples\": [\n 0,\n 1\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":37}],"source":["# to convert class labels from strings to integers\n","df['classlabel'] = df['classlabel'].map(class_mapping)\n","df"]},{"cell_type":"code","execution_count":38,"metadata":{"id":"Rfk1UdV2iE0O","executionInfo":{"status":"ok","timestamp":1726116706203,"user_tz":240,"elapsed":201,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/","height":160},"outputId":"4d8e496f-5603-4535-ae57-3d313fa24c78"},"outputs":[{"output_type":"execute_result","data":{"text/plain":[" color size price classlabel\n","0 green 1 10.1 class2\n","1 red 2 13.5 class1\n","2 blue 3 15.3 class2"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
colorsizepriceclasslabel
0green110.1class2
1red213.5class1
2blue315.3class2
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n"," \n"," \n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"df","summary":"{\n \"name\": \"df\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"color\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 3,\n \"samples\": [\n \"green\",\n \"red\",\n \"blue\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"size\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 1,\n \"min\": 1,\n \"max\": 3,\n \"num_unique_values\": 3,\n \"samples\": [\n 1,\n 2,\n 3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"price\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 2.640706976045115,\n \"min\": 10.1,\n \"max\": 15.3,\n \"num_unique_values\": 3,\n \"samples\": [\n 10.1,\n 13.5,\n 15.3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"classlabel\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 2,\n \"samples\": [\n \"class1\",\n \"class2\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":38}],"source":["# reverse the class label mapping\n","inv_class_mapping = {v: k for k, v in class_mapping.items()}\n","df['classlabel'] = df['classlabel'].map(inv_class_mapping)\n","df"]},{"cell_type":"code","source":["df['classlabel']"],"metadata":{"id":"_PSSE1Qm0gWk","executionInfo":{"status":"ok","timestamp":1726116708673,"user_tz":240,"elapsed":194,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/","height":193},"outputId":"73db5661-76f7-46b1-b8a5-8de543a77216"},"execution_count":39,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0 class2\n","1 class1\n","2 class2\n","Name: classlabel, dtype: object"],"text/html":["
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
classlabel
0class2
1class1
2class2
\n","

"]},"metadata":{},"execution_count":39}]},{"cell_type":"code","source":["df['classlabel'].values"],"metadata":{"id":"FY7J77qc0jxO","executionInfo":{"status":"ok","timestamp":1726116710570,"user_tz":240,"elapsed":151,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/"},"outputId":"467b0d7b-cd5b-411d-8a8e-88a612a94a9b"},"execution_count":40,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array(['class2', 'class1', 'class2'], dtype=object)"]},"metadata":{},"execution_count":40}]},{"cell_type":"code","source":["df"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":160},"id":"GdvV35em0yHo","executionInfo":{"status":"ok","timestamp":1726116711543,"user_tz":240,"elapsed":166,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"3caa1b36-19a6-4407-df13-730db774ea05"},"execution_count":41,"outputs":[{"output_type":"execute_result","data":{"text/plain":[" color size price classlabel\n","0 green 1 10.1 class2\n","1 red 2 13.5 class1\n","2 blue 3 15.3 class2"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
colorsizepriceclasslabel
0green110.1class2
1red213.5class1
2blue315.3class2
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n"," \n"," \n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"df","summary":"{\n \"name\": \"df\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"color\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 3,\n \"samples\": [\n \"green\",\n \"red\",\n \"blue\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"size\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 1,\n \"min\": 1,\n \"max\": 3,\n \"num_unique_values\": 3,\n \"samples\": [\n 1,\n 2,\n 3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"price\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 2.640706976045115,\n \"min\": 10.1,\n \"max\": 15.3,\n \"num_unique_values\": 3,\n \"samples\": [\n 10.1,\n 13.5,\n 15.3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"classlabel\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 2,\n \"samples\": [\n \"class1\",\n \"class2\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":41}]},{"cell_type":"code","execution_count":42,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"zv2gjD3UiE0O","executionInfo":{"status":"ok","timestamp":1726116724072,"user_tz":240,"elapsed":185,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"3b255a3c-aa1a-49be-e704-9bf7fbd661de"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([1, 0, 1])"]},"metadata":{},"execution_count":42}],"source":["from sklearn.preprocessing import LabelEncoder\n","\n","# Label encoding with sklearn's LabelEncoder\n","class_le = LabelEncoder()\n","y = class_le.fit_transform(df['classlabel'].values)\n","y"]},{"cell_type":"code","source":[],"metadata":{"id":"j3Piyaxc0iJJ","executionInfo":{"status":"ok","timestamp":1726116725679,"user_tz":240,"elapsed":184,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}}},"execution_count":42,"outputs":[]},{"cell_type":"code","execution_count":43,"metadata":{"id":"JWkqOSFhiE0O","executionInfo":{"status":"ok","timestamp":1726116726624,"user_tz":240,"elapsed":203,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/"},"outputId":"45214dde-33c8-4943-a4fc-977fddf6ae41"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["array(['class2', 'class1', 'class2'], dtype=object)"]},"metadata":{},"execution_count":43}],"source":["# reverse mapping\n","class_le.inverse_transform(y)"]},{"cell_type":"markdown","metadata":{"id":"ZVBRIPr0iE0O"},"source":["
\n","
"]},{"cell_type":"markdown","metadata":{"id":"HFs7rN22iE0O"},"source":["### Performing one-hot encoding on nominal features"]},{"cell_type":"code","source":["df"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":160},"id":"uQUIndYF3KIE","executionInfo":{"status":"ok","timestamp":1726116733173,"user_tz":240,"elapsed":154,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"7873d219-bb60-44f1-aa42-44c2819a3f35"},"execution_count":44,"outputs":[{"output_type":"execute_result","data":{"text/plain":[" color size price classlabel\n","0 green 1 10.1 class2\n","1 red 2 13.5 class1\n","2 blue 3 15.3 class2"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
colorsizepriceclasslabel
0green110.1class2
1red213.5class1
2blue315.3class2
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n"," \n"," \n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"df","summary":"{\n \"name\": \"df\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"color\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 3,\n \"samples\": [\n \"green\",\n \"red\",\n \"blue\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"size\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 1,\n \"min\": 1,\n \"max\": 3,\n \"num_unique_values\": 3,\n \"samples\": [\n 1,\n 2,\n 3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"price\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 2.640706976045115,\n \"min\": 10.1,\n \"max\": 15.3,\n \"num_unique_values\": 3,\n \"samples\": [\n 10.1,\n 13.5,\n 15.3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"classlabel\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 2,\n \"samples\": [\n \"class1\",\n \"class2\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":44}]},{"cell_type":"code","execution_count":45,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"iO8zUdFtiE0P","executionInfo":{"status":"ok","timestamp":1726116735169,"user_tz":240,"elapsed":184,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"742651c9-f589-43c4-ee4f-b229fee5c689"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[1, 1, 10.1],\n"," [2, 2, 13.5],\n"," [0, 3, 15.3]], dtype=object)"]},"metadata":{},"execution_count":45}],"source":["X = df[['color', 'size', 'price']].values\n","color_le = LabelEncoder()\n","X[:, 0] = color_le.fit_transform(X[:, 0])\n","X\n"]},{"cell_type":"code","source":["np.shape(X[:, 0])"],"metadata":{"id":"9Zthons25dAr","executionInfo":{"status":"ok","timestamp":1726116736677,"user_tz":240,"elapsed":191,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/"},"outputId":"b577b6b7-2367-4176-bb67-1de06e57779f"},"execution_count":46,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(3,)"]},"metadata":{},"execution_count":46}]},{"cell_type":"code","source":["np.shape(X[:, 0].reshape(-1, 1))"],"metadata":{"id":"nn33XXn85cVH","executionInfo":{"status":"ok","timestamp":1726116737804,"user_tz":240,"elapsed":183,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/"},"outputId":"848b245c-7538-404a-df0e-ca56e5504dde"},"execution_count":47,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(3, 1)"]},"metadata":{},"execution_count":47}]},{"cell_type":"code","execution_count":48,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"JqG2n6ROiE0P","executionInfo":{"status":"ok","timestamp":1726116739577,"user_tz":240,"elapsed":174,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"d287d91d-8c8c-471a-caa9-a33fa45ca126"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[0., 1., 0.],\n"," [0., 0., 1.],\n"," [1., 0., 0.]])"]},"metadata":{},"execution_count":48}],"source":["from sklearn.preprocessing import OneHotEncoder\n","\n","X = df[['color', 'size', 'price']].values\n","color_ohe = OneHotEncoder()\n","color_ohe.fit_transform(X[:, 0].reshape(-1, 1)).toarray()"]},{"cell_type":"code","source":["np.shape(color_ohe.fit_transform(X[:, 0].reshape(-1, 1)))"],"metadata":{"id":"m1495G1w9P_R","executionInfo":{"status":"ok","timestamp":1726116741681,"user_tz":240,"elapsed":170,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/"},"outputId":"8342f31c-fb9e-4b54-dad7-9a60b96345b6"},"execution_count":49,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(3, 3)"]},"metadata":{},"execution_count":49}]},{"cell_type":"code","source":["np.shape(color_ohe.fit_transform(X[:, 0].reshape(-1, 1)).toarray())"],"metadata":{"id":"38EvoyBV9T2n","executionInfo":{"status":"ok","timestamp":1726116742790,"user_tz":240,"elapsed":184,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/"},"outputId":"d04504cb-c91f-4d3b-d983-2fd7996f5798"},"execution_count":50,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(3, 3)"]},"metadata":{},"execution_count":50}]},{"cell_type":"code","execution_count":51,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"EvXtyx3oiE0P","executionInfo":{"status":"ok","timestamp":1726116743990,"user_tz":240,"elapsed":200,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"47f4c9b3-b914-4995-9e50-16b454c5ceec"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[ 0. , 1. , 0. , 1. , 10.1],\n"," [ 0. , 0. , 1. , 2. , 13.5],\n"," [ 1. , 0. , 0. , 3. , 15.3]])"]},"metadata":{},"execution_count":51}],"source":["from sklearn.compose import ColumnTransformer\n","\n","X = df[['color', 'size', 'price']].values\n","c_transf = ColumnTransformer([ ('onehot', OneHotEncoder(), [0]),\n"," ('nothing', 'passthrough', [1, 2])])\n","c_transf.fit_transform(X).astype(float)"]},{"cell_type":"code","source":["dh = pd.DataFrame([['green',10.2, 1],['blue',8.2, 1],['yellow',9.8, 0]])\n","dh.columns = ['colors','price','label']\n","dh"],"metadata":{"id":"rEWSROj0-yYW","executionInfo":{"status":"ok","timestamp":1726116745793,"user_tz":240,"elapsed":191,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/","height":160},"outputId":"6ee84584-7068-4ed0-cdc2-320114903d7b"},"execution_count":52,"outputs":[{"output_type":"execute_result","data":{"text/plain":[" colors price label\n","0 green 10.2 1\n","1 blue 8.2 1\n","2 yellow 9.8 0"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
colorspricelabel
0green10.21
1blue8.21
2yellow9.80
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n"," \n"," \n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"dh","summary":"{\n \"name\": \"dh\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"colors\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 3,\n \"samples\": [\n \"green\",\n \"blue\",\n \"yellow\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"price\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 1.0583005244258366,\n \"min\": 8.2,\n \"max\": 10.2,\n \"num_unique_values\": 3,\n \"samples\": [\n 10.2,\n 8.2,\n 9.8\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"label\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 0,\n \"min\": 0,\n \"max\": 1,\n \"num_unique_values\": 2,\n \"samples\": [\n 0,\n 1\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":52}]},{"cell_type":"code","source":["dh['colors']"],"metadata":{"id":"apmPuMn_Aaq5","executionInfo":{"status":"ok","timestamp":1726116747644,"user_tz":240,"elapsed":188,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/","height":193},"outputId":"ce102984-0d42-478d-b03d-6b097fbf4d05"},"execution_count":53,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0 green\n","1 blue\n","2 yellow\n","Name: colors, dtype: object"],"text/html":["
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
colors
0green
1blue
2yellow
\n","

"]},"metadata":{},"execution_count":53}]},{"cell_type":"code","source":["new_var = pd.get_dummies(dh[['colors','price','label']])\n"],"metadata":{"id":"erZ0Yi8__uGB","executionInfo":{"status":"ok","timestamp":1726116749956,"user_tz":240,"elapsed":195,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}}},"execution_count":54,"outputs":[]},{"cell_type":"code","source":["df\n"],"metadata":{"id":"Jp_-gMpfBCkX","executionInfo":{"status":"ok","timestamp":1726116752115,"user_tz":240,"elapsed":191,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"colab":{"base_uri":"https://localhost:8080/","height":160},"outputId":"7d9aec82-75a1-4c6a-a663-3fbc2e062aba"},"execution_count":55,"outputs":[{"output_type":"execute_result","data":{"text/plain":[" color size price classlabel\n","0 green 1 10.1 class2\n","1 red 2 13.5 class1\n","2 blue 3 15.3 class2"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
colorsizepriceclasslabel
0green110.1class2
1red213.5class1
2blue315.3class2
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n"," \n"," \n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"df","summary":"{\n \"name\": \"df\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"color\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 3,\n \"samples\": [\n \"green\",\n \"red\",\n \"blue\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"size\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 1,\n \"min\": 1,\n \"max\": 3,\n \"num_unique_values\": 3,\n \"samples\": [\n 1,\n 2,\n 3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"price\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 2.640706976045115,\n \"min\": 10.1,\n \"max\": 15.3,\n \"num_unique_values\": 3,\n \"samples\": [\n 10.1,\n 13.5,\n 15.3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"classlabel\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 2,\n \"samples\": [\n \"class1\",\n \"class2\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":55}]},{"cell_type":"code","execution_count":56,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":160},"id":"YDghXOhriE0P","executionInfo":{"status":"ok","timestamp":1726116754384,"user_tz":240,"elapsed":301,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"2b3ca6c9-7587-48a3-d70d-fed3184ddb0d"},"outputs":[{"output_type":"execute_result","data":{"text/plain":[" price size color_blue color_green color_red\n","0 10 1 0 1 0\n","1 13 2 0 0 1\n","2 15 3 1 0 0"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
pricesizecolor_bluecolor_greencolor_red
0101010
1132001
2153100
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","summary":"{\n \"name\": \"pd\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"price\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 2,\n \"min\": 10,\n \"max\": 15,\n \"num_unique_values\": 3,\n \"samples\": [\n 10,\n 13,\n 15\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"size\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 1,\n \"min\": 1,\n \"max\": 3,\n \"num_unique_values\": 3,\n \"samples\": [\n 1,\n 2,\n 3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"color_blue\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 0,\n \"min\": 0,\n \"max\": 1,\n \"num_unique_values\": 2,\n \"samples\": [\n 1,\n 0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"color_green\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 0,\n \"min\": 0,\n \"max\": 1,\n \"num_unique_values\": 2,\n \"samples\": [\n 0,\n 1\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"color_red\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 0,\n \"min\": 0,\n \"max\": 1,\n \"num_unique_values\": 2,\n \"samples\": [\n 1,\n 0\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":56}],"source":["# one-hot encoding via pandas\n","\n","pd.get_dummies(df[['price', 'color', 'size']]).astype(int)"]},{"cell_type":"code","execution_count":57,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":160},"id":"PMlwIufciE0P","executionInfo":{"status":"ok","timestamp":1726116755889,"user_tz":240,"elapsed":184,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"f869cf04-7c43-40c3-a9f0-fafe1b3d93ca"},"outputs":[{"output_type":"execute_result","data":{"text/plain":[" price size color_green color_red\n","0 10.1 1 True False\n","1 13.5 2 False True\n","2 15.3 3 False False"],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
pricesizecolor_greencolor_red
010.11TrueFalse
113.52FalseTrue
215.33FalseFalse
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","summary":"{\n \"name\": \"pd\",\n \"rows\": 3,\n \"fields\": [\n {\n \"column\": \"price\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 2.640706976045115,\n \"min\": 10.1,\n \"max\": 15.3,\n \"num_unique_values\": 3,\n \"samples\": [\n 10.1,\n 13.5,\n 15.3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"size\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 1,\n \"min\": 1,\n \"max\": 3,\n \"num_unique_values\": 3,\n \"samples\": [\n 1,\n 2,\n 3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"color_green\",\n \"properties\": {\n \"dtype\": \"boolean\",\n \"num_unique_values\": 2,\n \"samples\": [\n false,\n true\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"color_red\",\n \"properties\": {\n \"dtype\": \"boolean\",\n \"num_unique_values\": 2,\n \"samples\": [\n true,\n false\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":57}],"source":["# multicollinearity in get_dummies\n","\n","pd.get_dummies(df[['price', 'color', 'size']], drop_first=True)"]},{"cell_type":"code","execution_count":58,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"kyNigmVriE0P","executionInfo":{"status":"ok","timestamp":1726116757296,"user_tz":240,"elapsed":187,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"da88b4c2-b298-41ac-a797-60f8238025c6"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[ 1. , 0. , 1. , 10.1],\n"," [ 0. , 1. , 2. , 13.5],\n"," [ 0. , 0. , 3. , 15.3]])"]},"metadata":{},"execution_count":58}],"source":["# multicollinearity guard for the OneHotEncoder\n","\n","color_ohe = OneHotEncoder(categories='auto', drop='first')\n","c_transf = ColumnTransformer([ ('onehot', color_ohe, [0]),\n"," ('nothing', 'passthrough', [1, 2])])\n","c_transf.fit_transform(X).astype(float)"]},{"cell_type":"markdown","metadata":{"id":"BoCgI3lDiE0P"},"source":["
\n","
"]},{"cell_type":"markdown","metadata":{"id":"UvPxkvEciE0Q"},"source":["
\n","
"]},{"cell_type":"markdown","metadata":{"id":"eBmWjvEQiE0Q"},"source":["### Partitioning a dataset into a separate training and test set"]},{"cell_type":"code","execution_count":59,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":271},"id":"uYlIYVsGiE0Q","executionInfo":{"status":"ok","timestamp":1726116798464,"user_tz":240,"elapsed":658,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"cfd450f9-150a-44e6-d89c-2253d49b1e2f"},"outputs":[{"output_type":"stream","name":"stdout","text":["Class labels [1 2 3]\n"]},{"output_type":"execute_result","data":{"text/plain":[" Class label Alcohol Malic acid Ash Alcalinity of ash Magnesium \\\n","0 1 14.23 1.71 2.43 15.6 127 \n","1 1 13.20 1.78 2.14 11.2 100 \n","2 1 13.16 2.36 2.67 18.6 101 \n","3 1 14.37 1.95 2.50 16.8 113 \n","4 1 13.24 2.59 2.87 21.0 118 \n","\n"," Total phenols Flavanoids Nonflavanoid phenols Proanthocyanins \\\n","0 2.80 3.06 0.28 2.29 \n","1 2.65 2.76 0.26 1.28 \n","2 2.80 3.24 0.30 2.81 \n","3 3.85 3.49 0.24 2.18 \n","4 2.80 2.69 0.39 1.82 \n","\n"," Color intensity Hue OD280/OD315 of diluted wines Proline \n","0 5.64 1.04 3.92 1065 \n","1 4.38 1.05 3.40 1050 \n","2 5.68 1.03 3.17 1185 \n","3 7.80 0.86 3.45 1480 \n","4 4.32 1.04 2.93 735 "],"text/html":["\n","
\n","
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
Class labelAlcoholMalic acidAshAlcalinity of ashMagnesiumTotal phenolsFlavanoidsNonflavanoid phenolsProanthocyaninsColor intensityHueOD280/OD315 of diluted winesProline
0114.231.712.4315.61272.803.060.282.295.641.043.921065
1113.201.782.1411.21002.652.760.261.284.381.053.401050
2113.162.362.6718.61012.803.240.302.815.681.033.171185
3114.371.952.5016.81133.853.490.242.187.800.863.451480
4113.242.592.8721.01182.802.690.391.824.321.042.93735
\n","
\n","
\n","\n","
\n"," \n","\n"," \n","\n"," \n","
\n","\n","\n","
\n"," \n","\n","\n","\n"," \n","
\n","\n","
\n","
\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"df_wine","summary":"{\n \"name\": \"df_wine\",\n \"rows\": 178,\n \"fields\": [\n {\n \"column\": \"Class label\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 0,\n \"min\": 1,\n \"max\": 3,\n \"num_unique_values\": 3,\n \"samples\": [\n 1,\n 2,\n 3\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"Alcohol\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 0.8118265380058575,\n \"min\": 11.03,\n \"max\": 14.83,\n \"num_unique_values\": 126,\n \"samples\": [\n 11.62,\n 13.64,\n 13.69\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"Malic acid\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 1.1171460976144627,\n \"min\": 0.74,\n \"max\": 5.8,\n \"num_unique_values\": 133,\n \"samples\": [\n 1.21,\n 2.83,\n 1.8\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"Ash\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 0.27434400906081485,\n \"min\": 1.36,\n \"max\": 3.23,\n \"num_unique_values\": 79,\n \"samples\": [\n 2.31,\n 2.43,\n 2.52\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"Alcalinity of ash\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 3.339563767173505,\n \"min\": 10.6,\n \"max\": 30.0,\n \"num_unique_values\": 63,\n \"samples\": [\n 25.5,\n 28.5,\n 15.6\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"Magnesium\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 14,\n \"min\": 70,\n \"max\": 162,\n \"num_unique_values\": 53,\n \"samples\": [\n 126,\n 85,\n 162\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"Total phenols\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 0.6258510488339893,\n \"min\": 0.98,\n \"max\": 3.88,\n \"num_unique_values\": 97,\n \"samples\": [\n 1.68,\n 2.11,\n 1.35\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"Flavanoids\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 0.9988586850169467,\n \"min\": 0.34,\n \"max\": 5.08,\n \"num_unique_values\": 132,\n \"samples\": [\n 3.18,\n 2.5,\n 3.17\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"Nonflavanoid phenols\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 0.12445334029667937,\n \"min\": 0.13,\n \"max\": 0.66,\n \"num_unique_values\": 39,\n \"samples\": [\n 0.58,\n 0.41,\n 0.39\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"Proanthocyanins\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 0.5723588626747613,\n \"min\": 0.41,\n \"max\": 3.58,\n \"num_unique_values\": 101,\n \"samples\": [\n 0.75,\n 1.77,\n 1.42\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"Color intensity\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 2.318285871822413,\n \"min\": 1.28,\n \"max\": 13.0,\n \"num_unique_values\": 132,\n \"samples\": [\n 2.95,\n 3.3,\n 5.1\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"Hue\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 0.22857156582982338,\n \"min\": 0.48,\n \"max\": 1.71,\n \"num_unique_values\": 78,\n \"samples\": [\n 1.22,\n 1.04,\n 1.45\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"OD280/OD315 of diluted wines\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 0.7099904287650504,\n \"min\": 1.27,\n \"max\": 4.0,\n \"num_unique_values\": 122,\n \"samples\": [\n 4.0,\n 1.82,\n 1.59\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"Proline\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 314,\n \"min\": 278,\n \"max\": 1680,\n \"num_unique_values\": 121,\n \"samples\": [\n 1375,\n 1270,\n 735\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}"}},"metadata":{},"execution_count":59}],"source":["df_wine = pd.read_csv('https://archive.ics.uci.edu/'\n"," 'ml/machine-learning-databases/wine/wine.data',\n"," header=None)\n","\n","df_wine.columns = ['Class label', 'Alcohol', 'Malic acid', 'Ash',\n"," 'Alcalinity of ash', 'Magnesium', 'Total phenols',\n"," 'Flavanoids', 'Nonflavanoid phenols', 'Proanthocyanins',\n"," 'Color intensity', 'Hue', 'OD280/OD315 of diluted wines',\n"," 'Proline']\n","\n","print('Class labels', np.unique(df_wine['Class label']))\n","df_wine.head()"]},{"cell_type":"code","execution_count":60,"metadata":{"id":"aTW2BCB6iE0R","executionInfo":{"status":"ok","timestamp":1726116804587,"user_tz":240,"elapsed":181,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}}},"outputs":[],"source":["from sklearn.model_selection import train_test_split\n","\n","X, y = df_wine.iloc[:, 1:].values, df_wine.iloc[:, 0].values\n","\n","X_train, X_test, y_train, y_test =\\\n"," train_test_split(X, y,\n"," test_size=0.3,\n"," random_state=0,\n"," stratify=y)"]},{"cell_type":"markdown","metadata":{"id":"x5dYKuCwiE0R"},"source":["
\n","
"]},{"cell_type":"markdown","metadata":{"id":"zj8VZzdsiE0R"},"source":["### Bringing features onto the same scale"]},{"cell_type":"code","execution_count":61,"metadata":{"id":"sSekDhyKiE0R","executionInfo":{"status":"ok","timestamp":1726116815228,"user_tz":240,"elapsed":153,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}}},"outputs":[],"source":["from sklearn.preprocessing import MinMaxScaler\n","\n","mms = MinMaxScaler()\n","X_train_norm = mms.fit_transform(X_train)\n","X_test_norm = mms.transform(X_test)"]},{"cell_type":"code","execution_count":62,"metadata":{"id":"u9F-CVrLiE0R","executionInfo":{"status":"ok","timestamp":1726116816378,"user_tz":240,"elapsed":176,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}}},"outputs":[],"source":["from sklearn.preprocessing import StandardScaler\n","\n","stdsc = StandardScaler()\n","X_train_std = stdsc.fit_transform(X_train)\n","X_test_std = stdsc.transform(X_test)"]},{"cell_type":"markdown","metadata":{"id":"CNyu_3v0iE0R"},"source":["A visual example:"]},{"cell_type":"code","execution_count":63,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"GkQuQFPeiE0R","executionInfo":{"status":"ok","timestamp":1726116818022,"user_tz":240,"elapsed":180,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"6cceff87-a393-4eae-f5bc-4fade89c9bef"},"outputs":[{"output_type":"stream","name":"stdout","text":["standardized: [-1.46385011 -0.87831007 -0.29277002 0.29277002 0.87831007 1.46385011]\n","normalized: [0. 0.2 0.4 0.6 0.8 1. ]\n"]}],"source":["ex = np.array([0, 1, 2, 3, 4, 5])\n","\n","print('standardized:', (ex - ex.mean()) / ex.std())\n","\n","# Please note that pandas uses ddof=1 (sample standard deviation)\n","# by default, whereas NumPy's std method and the StandardScaler\n","# uses ddof=0 (population standard deviation)\n","\n","# normalize\n","print('normalized:', (ex - ex.min()) / (ex.max() - ex.min()))"]},{"cell_type":"markdown","metadata":{"id":"c1yAKvMoiE0S"},"source":["
\n","
"]},{"cell_type":"markdown","metadata":{"id":"7BQlKjFuiE0S"},"source":["### Selecting meaningful features"]},{"cell_type":"markdown","metadata":{"id":"0N4KHoJ9iE0S"},"source":["---- Sparse solutions with L1-regularization"]},{"cell_type":"markdown","metadata":{"id":"Z-yaHKI5iE0T"},"source":["For regularized models in scikit-learn that support L1 regularization, we can simply set the `penalty` parameter to `'l1'` to obtain a sparse solution:"]},{"cell_type":"code","execution_count":64,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":93},"id":"O669U25NiE0T","executionInfo":{"status":"ok","timestamp":1726116869603,"user_tz":240,"elapsed":164,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"239f4562-9ed4-42ca-9d52-eabe93752bc6"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["LogisticRegression(penalty='l1')"],"text/html":["
LogisticRegression(penalty='l1')
In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.
"]},"metadata":{},"execution_count":64}],"source":["from sklearn.linear_model import LogisticRegression\n","\n","LogisticRegression(penalty='l1')"]},{"cell_type":"markdown","metadata":{"id":"X59Bhyy9iE0T"},"source":["Applied to the standardized Wine data ..."]},{"cell_type":"code","execution_count":65,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"iFDNC3L8iE0T","executionInfo":{"status":"ok","timestamp":1726116874609,"user_tz":240,"elapsed":164,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"b4cbd77f-ea5d-4c9d-9203-626c4fd6ce94"},"outputs":[{"output_type":"stream","name":"stdout","text":["Training accuracy: 1.0\n","Test accuracy: 1.0\n"]}],"source":["from sklearn.linear_model import LogisticRegression\n","\n","lr = LogisticRegression(penalty='l1', C=1.0, solver='liblinear', multi_class='ovr')\n","# Note that C=1.0 is the default. You can increase\n","# or decrease it to make the regulariztion effect\n","# stronger or weaker, respectively.\n","lr.fit(X_train_std, y_train)\n","print('Training accuracy:', lr.score(X_train_std, y_train))\n","print('Test accuracy:', lr.score(X_test_std, y_test))"]},{"cell_type":"code","execution_count":66,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"5v_Dk6KgiE0T","executionInfo":{"status":"ok","timestamp":1726116876354,"user_tz":240,"elapsed":154,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"8fb35cd6-2f26-464d-e8ec-6597057c93d9"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([-1.26352457, -1.21576785, -2.3711671 ])"]},"metadata":{},"execution_count":66}],"source":["lr.intercept_"]},{"cell_type":"code","execution_count":67,"metadata":{"id":"ujLtZU9LiE0T","executionInfo":{"status":"ok","timestamp":1726116876898,"user_tz":240,"elapsed":2,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}}},"outputs":[],"source":["np.set_printoptions(8)"]},{"cell_type":"code","execution_count":68,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"4TXhFAkwiE0T","executionInfo":{"status":"ok","timestamp":1726116877465,"user_tz":240,"elapsed":3,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"67296a31-93e2-4740-9c02-1118629149c8"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["(23,)"]},"metadata":{},"execution_count":68}],"source":["lr.coef_[lr.coef_!=0].shape"]},{"cell_type":"code","execution_count":69,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"wpx7vjS2iE0U","executionInfo":{"status":"ok","timestamp":1726116878806,"user_tz":240,"elapsed":292,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"6b412bbe-69b7-415b-da15-b0eb0bbb8726"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[ 1.24574039, 0.18046984, 0.74484363, -1.16251576, 0. ,\n"," 0. , 1.16495206, 0. , 0. , 0. ,\n"," 0. , 0.55194572, 2.50968308],\n"," [-1.53727681, -0.38718475, -0.99524098, 0.36476702, -0.05956256,\n"," 0. , 0.6679675 , 0. , 0. , -1.93383967,\n"," 1.23411639, 0. , -2.23183041],\n"," [ 0.13539093, 0.16972595, 0.35779691, 0. , 0. ,\n"," 0. , -2.43329298, 0. , 0. , 1.56172158,\n"," -0.81756997, -0.49748177, 0. ]])"]},"metadata":{},"execution_count":69}],"source":["lr.coef_"]},{"cell_type":"code","source":["np.shape(lr.coef_)"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"f6enq1Tp3bN-","executionInfo":{"status":"ok","timestamp":1726116880341,"user_tz":240,"elapsed":192,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"d5d9abdd-c346-4be3-cba3-856aeffda214"},"execution_count":70,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(3, 13)"]},"metadata":{},"execution_count":70}]},{"cell_type":"code","source":["lr.coef_[1]"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Qw1-sASD3vcC","executionInfo":{"status":"ok","timestamp":1726116882732,"user_tz":240,"elapsed":3,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"a9412b46-c740-4ae3-e10d-e841d28b34a4"},"execution_count":71,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([-1.53727681, -0.38718475, -0.99524098, 0.36476702, -0.05956256,\n"," 0. , 0.6679675 , 0. , 0. , -1.93383967,\n"," 1.23411639, 0. , -2.23183041])"]},"metadata":{},"execution_count":71}]},{"cell_type":"code","execution_count":72,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":710},"id":"OpYrMJY2iE0U","executionInfo":{"status":"ok","timestamp":1726116888511,"user_tz":240,"elapsed":1201,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"3b4a90c7-4780-404d-b98b-395f8fdb6350"},"outputs":[{"output_type":"stream","name":"stdout","text":["(10, 13)\n","0 blue\n","1 green\n","2 red\n","3 cyan\n","4 magenta\n","5 yellow\n","6 black\n","7 pink\n","8 lightgreen\n","9 lightblue\n","10 gray\n","11 indigo\n","12 orange\n"]},{"output_type":"display_data","data":{"text/plain":["
"],"image/png":"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\n"},"metadata":{}}],"source":["import matplotlib.pyplot as plt\n","\n","fig = plt.figure()\n","ax = plt.subplot(111)\n","\n","colors = ['blue', 'green', 'red', 'cyan',\n"," 'magenta', 'yellow', 'black',\n"," 'pink', 'lightgreen', 'lightblue',\n"," 'gray', 'indigo', 'orange']\n","\n","weights, params = [], []\n","for c in np.arange(-4., 6.):\n"," lr = LogisticRegression(penalty='l1', C=10.**c, solver='liblinear',\n"," multi_class='ovr', random_state=0)\n"," lr.fit(X_train_std, y_train)\n"," weights.append(lr.coef_[1])\n"," params.append(10**c)\n","\n","weights = np.array(weights)\n","\n","print(np.shape(weights)) #dims of weights (10,13)\n","\n","# you want to show how each weight (corresponding to a feature) changes vs regularization intensity\n","\n","for column, color in zip(range(weights.shape[1]), colors):\n"," print(column,color)\n"," plt.plot(params, weights[:, column],\n"," label=df_wine.columns[column + 1],\n"," color=color)\n","plt.axhline(0, color='black', linestyle='--', linewidth=3)\n","plt.xlim([10**(-5), 10**5])\n","plt.ylabel('Weight coefficient')\n","plt.xlabel('C (inverse regularization strength)')\n","plt.xscale('log')\n","plt.legend(loc='upper left')\n","ax.legend(loc='upper center',\n"," bbox_to_anchor=(1.38, 1.03),\n"," ncol=1, fancybox=True)\n","\n","#plt.savefig('figures/04_08.png', dpi=300,\n","# bbox_inches='tight', pad_inches=0.2)\n","\n","plt.show()"]},{"cell_type":"markdown","metadata":{"id":"8GTlML9liE0b"},"source":["
\n","
"]},{"cell_type":"markdown","metadata":{"id":"_QArkxCmiE06"},"source":["
\n","
"]},{"cell_type":"markdown","metadata":{"id":"hrjfu4y4iE06"},"source":["### Assessing feature importance with Random Forests"]},{"cell_type":"code","execution_count":76,"metadata":{"id":"gZETDYdliE06","outputId":"5c9ce72a-008e-4eb0-d5a8-1e0fb2ba23de","colab":{"base_uri":"https://localhost:8080/","height":723},"executionInfo":{"status":"ok","timestamp":1726116940284,"user_tz":240,"elapsed":1752,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}}},"outputs":[{"output_type":"stream","name":"stdout","text":[" 1) Proline 0.185453\n"," 2) Flavanoids 0.174751\n"," 3) Color intensity 0.143920\n"," 4) OD280/OD315 of diluted wines 0.136162\n"," 5) Alcohol 0.118529\n"," 6) Hue 0.058739\n"," 7) Total phenols 0.050872\n"," 8) Magnesium 0.031357\n"," 9) Malic acid 0.025648\n","10) Proanthocyanins 0.025570\n","11) Alcalinity of ash 0.022366\n","12) Nonflavanoid phenols 0.013354\n","13) Ash 0.013279\n"]},{"output_type":"display_data","data":{"text/plain":["
"],"image/png":"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\n"},"metadata":{}}],"source":["from sklearn.ensemble import RandomForestClassifier\n","\n","feat_labels = df_wine.columns[1:]\n","\n","forest = RandomForestClassifier(n_estimators=500,\n"," random_state=1)\n","\n","forest.fit(X_train, y_train)\n","importances = forest.feature_importances_\n","\n","indices = np.argsort(importances)[::-1]\n","\n","for f in range(X_train.shape[1]):\n"," print(\"%2d) %-*s %f\" % (f + 1, 30,\n"," feat_labels[indices[f]],\n"," importances[indices[f]]))\n","\n","plt.title('Feature importance')\n","plt.bar(range(X_train.shape[1]),\n"," importances[indices],\n"," align='center')\n","\n","plt.xticks(range(X_train.shape[1]),\n"," feat_labels[indices], rotation=90)\n","plt.xlim([-1, X_train.shape[1]])\n","plt.tight_layout()\n","# plt.savefig('figures/04_10.png', dpi=300)\n","plt.show()"]},{"cell_type":"code","execution_count":77,"metadata":{"id":"mUW0vq4SiE06","outputId":"57f529a9-c0cd-4e69-ff85-afb770be40bf","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1726116942847,"user_tz":240,"elapsed":178,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}}},"outputs":[{"output_type":"stream","name":"stdout","text":["Number of features that meet this threshold criterion: 5\n"]}],"source":["from sklearn.feature_selection import SelectFromModel\n","\n","sfm = SelectFromModel(forest, threshold=0.1, prefit=True)\n","X_selected = sfm.transform(X_train)\n","print('Number of features that meet this threshold criterion:',\n"," X_selected.shape[1])"]},{"cell_type":"markdown","metadata":{"id":"8qlpoBk9iE07"},"source":["Now, let's print the 3 features that met the threshold criterion for feature selection that we set earlier (note that this code snippet does not appear in the actual book but was added to this notebook later for illustrative purposes):"]},{"cell_type":"code","execution_count":78,"metadata":{"id":"m-EELdmXiE07","outputId":"588a8c8e-9d5b-485e-91a4-aa73bd73c5fc","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1726116944189,"user_tz":240,"elapsed":188,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}}},"outputs":[{"output_type":"stream","name":"stdout","text":[" 1) Proline 0.185453\n"," 2) Flavanoids 0.174751\n"," 3) Color intensity 0.143920\n"," 4) OD280/OD315 of diluted wines 0.136162\n"," 5) Alcohol 0.118529\n"]}],"source":["for f in range(X_selected.shape[1]):\n"," print(\"%2d) %-*s %f\" % (f + 1, 30,\n"," feat_labels[indices[f]],\n"," importances[indices[f]]))"]},{"cell_type":"markdown","metadata":{"id":"u1FIfSr-iE07"},"source":["
\n","
"]},{"cell_type":"markdown","source":["---- Example in Python of Feature Importance for Regression problems"],"metadata":{"id":"BQ6tm0TRNwNO"}},{"cell_type":"code","source":["from sklearn.ensemble import RandomForestRegressor\n","from sklearn.datasets import fetch_california_housing\n","\n","# Load sample data\n","data = fetch_california_housing()\n","X, y = data.data, data.target\n","\n","# Fit Random Forest model\n","model = RandomForestRegressor(n_estimators=100, random_state=42)\n","model.fit(X, y)\n","\n","# Print feature importance\n","feature_importance = model.feature_importances_\n","for i, imp in enumerate(feature_importance):\n"," print(f\"Feature {data.feature_names[i]}: {imp}\")\n"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"kH0t6zJjNv0P","executionInfo":{"status":"ok","timestamp":1726116986741,"user_tz":240,"elapsed":25300,"user":{"displayName":"Cristiano Fanelli","userId":"17509928347219701270"}},"outputId":"89d160c4-beb7-460b-eac7-d76efdcd2996"},"execution_count":79,"outputs":[{"output_type":"stream","name":"stdout","text":["Feature MedInc: 0.5200367196529164\n","Feature HouseAge: 0.05296357881747684\n","Feature AveRooms: 0.04451309296326938\n","Feature AveBedrms: 0.029298856378707893\n","Feature Population: 0.03123174949895071\n","Feature AveOccup: 0.13640641507927073\n","Feature Latitude: 0.09285575343954347\n","Feature Longitude: 0.09269383416986465\n"]}]},{"cell_type":"code","source":[],"metadata":{"id":"2_DBkD3eN8ZS"},"execution_count":null,"outputs":[]}],"metadata":{"anaconda-cloud":{},"kernelspec":{"display_name":"Python 3 (ipykernel)","language":"python","name":"python3"},"language_info":{"codemirror_mode":{"name":"ipython","version":3},"file_extension":".py","mimetype":"text/x-python","name":"python","nbconvert_exporter":"python","pygments_lexer":"ipython3","version":"3.8.12"},"toc":{"nav_menu":{},"number_sections":true,"sideBar":true,"skip_h1_title":false,"title_cell":"Table of Contents","title_sidebar":"Contents","toc_cell":false,"toc_position":{},"toc_section_display":true,"toc_window_display":false},"colab":{"provenance":[]}},"nbformat":4,"nbformat_minor":0}