From 22af266e65bb6511c496bb0fb4450ba0c439255a Mon Sep 17 00:00:00 2001 From: benjas <909336740@qq.com> Date: Fri, 18 Dec 2020 10:33:21 +0800 Subject: [PATCH] =?UTF-8?q?Create=20=E5=B8=B8=E7=94=A8=E7=89=B9=E5=BE=81?= =?UTF-8?q?=E6=9E=84=E9=80=A0=E6=96=B9=E6=B3=95-checkpoint.ipynb?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .../常用特征构造方法-checkpoint.ipynb | 2765 +++++++++++++++++ 1 file changed, 2765 insertions(+) create mode 100644 机器学习竞赛实战_优胜解决方案/常用特征构建方法/.ipynb_checkpoints/常用特征构造方法-checkpoint.ipynb diff --git a/机器学习竞赛实战_优胜解决方案/常用特征构建方法/.ipynb_checkpoints/常用特征构造方法-checkpoint.ipynb b/机器学习竞赛实战_优胜解决方案/常用特征构建方法/.ipynb_checkpoints/常用特征构造方法-checkpoint.ipynb new file mode 100644 index 0000000..0b142fa --- /dev/null +++ b/机器学习竞赛实战_优胜解决方案/常用特征构建方法/.ipynb_checkpoints/常用特征构造方法-checkpoint.ipynb @@ -0,0 +1,2765 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## 离散值处理" + ] + }, + { + "cell_type": "code", + "execution_count": 1, + "metadata": {}, + "outputs": [], + "source": [ + "import pandas as pd\n", + "import numpy as np\n", + "import warnings # 忽略普通警告，不打印太多东西\n", + "warnings.filterwarnings('ignore')" + ] + }, + { + "cell_type": "code", + "execution_count": 2, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
NamePlatformYearGenrePublisher
1Super Mario Bros.NES1985.0PlatformNintendo
2Mario Kart WiiWii2008.0RacingNintendo
3Wii Sports ResortWii2009.0SportsNintendo
4Pokemon Red/Pokemon BlueGB1996.0Role-PlayingNintendo
5TetrisGB1989.0PuzzleNintendo
6New Super Mario Bros.DS2006.0PlatformNintendo
\n", + "
" + ], + "text/plain": [ + " Name Platform Year Genre Publisher\n", + "1 Super Mario Bros. NES 1985.0 Platform Nintendo\n", + "2 Mario Kart Wii Wii 2008.0 Racing Nintendo\n", + "3 Wii Sports Resort Wii 2009.0 Sports Nintendo\n", + "4 Pokemon Red/Pokemon Blue GB 1996.0 Role-Playing Nintendo\n", + "5 Tetris GB 1989.0 Puzzle Nintendo\n", + "6 New Super Mario Bros. DS 2006.0 Platform Nintendo" + ] + }, + "execution_count": 2, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "vg_df = pd.read_csv('data/vgsales.csv', encoding='ISO-8859-1')\n", + "vg_df[['Name', 'Platform', 'Year', 'Genre', 'Publisher']].iloc[1:7]" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "机器无法识别字符串类型数据，需要做处理" + ] + }, + { + "cell_type": "code", + "execution_count": 3, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "array(['Action', 'Adventure', 'Fighting', 'Misc', 'Platform', 'Puzzle',\n", + " 'Racing', 'Role-Playing', 'Shooter', 'Simulation', 'Sports',\n", + " 'Strategy'], dtype=object)" + ] + }, + "execution_count": 3, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "genres = np.unique(vg_df['Genre'])\n", + "genres # 不同的字符串并不多" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## LabelEncoder" + ] + }, + { + "cell_type": "code", + "execution_count": 4, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "{0: 'Action',\n", + " 1: 'Adventure',\n", + " 2: 'Fighting',\n", + " 3: 'Misc',\n", + " 4: 'Platform',\n", + " 5: 'Puzzle',\n", + " 6: 'Racing',\n", + " 7: 'Role-Playing',\n", + " 8: 'Shooter',\n", + " 9: 'Simulation',\n", + " 10: 'Sports',\n", + " 11: 'Strategy'}" + ] + }, + "execution_count": 4, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "from sklearn.preprocessing import LabelEncoder\n", + "\n", + "gle = LabelEncoder() # 实例化\n", + "genre_labels = gle.fit_transform(vg_df['Genre']) # 转换需要离散值的一列\n", + "genre_mappings = {index: label for index, label in enumerate(gle.classes_)}\n", + "genre_mappings # 映射成数值" + ] + }, + { + "cell_type": "code", + "execution_count": 5, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
NamePlatformYearGenreGenreLabel
1Super Mario Bros.NES1985.0Platform4
2Mario Kart WiiWii2008.0Racing6
3Wii Sports ResortWii2009.0Sports10
4Pokemon Red/Pokemon BlueGB1996.0Role-Playing7
5TetrisGB1989.0Puzzle5
6New Super Mario Bros.DS2006.0Platform4
\n", + "
" + ], + "text/plain": [ + " Name Platform Year Genre GenreLabel\n", + "1 Super Mario Bros. NES 1985.0 Platform 4\n", + "2 Mario Kart Wii Wii 2008.0 Racing 6\n", + "3 Wii Sports Resort Wii 2009.0 Sports 10\n", + "4 Pokemon Red/Pokemon Blue GB 1996.0 Role-Playing 7\n", + "5 Tetris GB 1989.0 Puzzle 5\n", + "6 New Super Mario Bros. DS 2006.0 Platform 4" + ] + }, + "execution_count": 5, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "vg_df['GenreLabel'] = genre_labels # 赋值到一列\n", + "vg_df[['Name', 'Platform', 'Year', 'Genre', 'GenreLabel']].iloc[1:7]" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## Map\n", + "自己建一个字典" + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "{'Action': 0,\n", + " 'Adventure': 1,\n", + " 'Fighting': 2,\n", + " 'Misc': 3,\n", + " 'Platform': 4,\n", + " 'Puzzle': 5,\n", + " 'Racing': 6,\n", + " 'Role-Playing': 7,\n", + " 'Shooter': 8,\n", + " 'Simulation': 9,\n", + " 'Sports': 10,\n", + " 'Strategy': 11}" + ] + }, + "execution_count": 6, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "gen_ord_map = {label:index for index, label in enumerate(gle.classes_)}\n", + "gen_ord_map" + ] + }, + { + "cell_type": "code", + "execution_count": 7, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
NameGenreGenreLabelGenreMap
1Super Mario Bros.Platform44
2Mario Kart WiiRacing66
3Wii Sports ResortSports1010
4Pokemon Red/Pokemon BlueRole-Playing77
5TetrisPuzzle55
6New Super Mario Bros.Platform44
\n", + "
" + ], + "text/plain": [ + " Name Genre GenreLabel GenreMap\n", + "1 Super Mario Bros. Platform 4 4\n", + "2 Mario Kart Wii Racing 6 6\n", + "3 Wii Sports Resort Sports 10 10\n", + "4 Pokemon Red/Pokemon Blue Role-Playing 7 7\n", + "5 Tetris Puzzle 5 5\n", + "6 New Super Mario Bros. Platform 4 4" + ] + }, + "execution_count": 7, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "vg_df['GenreMap'] = vg_df['Genre'].map(gen_ord_map)\n", + "vg_df[['Name', 'Genre', 'GenreLabel', 'GenreMap']].iloc[1:7] # 结果呈现我们设置的map" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## One-Hot Encoder\n", + "对于离散型特征，基于树的方法是不需要使用one-hot编码的，例如随机森林等。基于距离的模型，都是要使用one-hot编码，例如神经网络等。" + ] + }, + { + "cell_type": "code", + "execution_count": 8, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "array([[0., 0., 0., ..., 0., 1., 0.],\n", + " [0., 0., 0., ..., 0., 0., 0.],\n", + " [0., 0., 0., ..., 0., 0., 0.],\n", + " ...,\n", + " [0., 0., 0., ..., 0., 0., 0.],\n", + " [0., 0., 0., ..., 0., 0., 0.],\n", + " [0., 0., 0., ..., 0., 0., 0.]])" + ] + }, + "execution_count": 8, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "from sklearn.preprocessing import OneHotEncoder\n", + "# 获取onehot后的结果，将字符串变成多列的0/1值，有则为1，无则为0\n", + "gen_ohe = OneHotEncoder()\n", + "gen_feature_arr = gen_ohe.fit_transform(vg_df[['GenreLabel']]).toarray()\n", + "gen_feature_arr" + ] + }, + { + "cell_type": "code", + "execution_count": 9, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
ActionAdventureFightingMiscPlatformPuzzleRacingRole-PlayingShooterSimulationSportsStrategy
00.00.00.00.00.00.00.00.00.00.01.00.0
10.00.00.00.01.00.00.00.00.00.00.00.0
20.00.00.00.00.00.01.00.00.00.00.00.0
30.00.00.00.00.00.00.00.00.00.01.00.0
40.00.00.00.00.00.00.01.00.00.00.00.0
\n", + "
" + ], + "text/plain": [ + " Action Adventure Fighting Misc Platform Puzzle Racing Role-Playing \\\n", + "0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 \n", + "1 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 \n", + "2 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 \n", + "3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 \n", + "4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 \n", + "\n", + " Shooter Simulation Sports Strategy \n", + "0 0.0 0.0 1.0 0.0 \n", + "1 0.0 0.0 0.0 0.0 \n", + "2 0.0 0.0 0.0 0.0 \n", + "3 0.0 0.0 1.0 0.0 \n", + "4 0.0 0.0 0.0 0.0 " + ] + }, + "execution_count": 9, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "genres = np.unique(vg_df['Genre']) # 获取全部不同的字符串\n", + "gen_features = pd.DataFrame(gen_feature_arr, columns=genres) # 将字符串作为列，合并onehot数据\n", + "gen_features.head()" + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "metadata": { + "scrolled": true + }, + "outputs": [ + { + "data": { + "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", + "
NameGenre
0Wii SportsSports
1Super Mario Bros.Platform
2Mario Kart WiiRacing
3Wii Sports ResortSports
4Pokemon Red/Pokemon BlueRole-Playing
\n", + "
" + ], + "text/plain": [ + " Name Genre\n", + "0 Wii Sports Sports\n", + "1 Super Mario Bros. Platform\n", + "2 Mario Kart Wii Racing\n", + "3 Wii Sports Resort Sports\n", + "4 Pokemon Red/Pokemon Blue Role-Playing" + ] + }, + "execution_count": 10, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "# 拿出两列原本的数据，实际场景中是全部数据合并，这里是为了查看方便\n", + "vg_df_2 = vg_df[['Name', 'Genre']]\n", + "vg_df_2.head()" + ] + }, + { + "cell_type": "code", + "execution_count": 11, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
NameGenreActionAdventureFightingMiscPlatformPuzzleRacingRole-PlayingShooterSimulationSportsStrategy
0Wii SportsSports0.00.00.00.00.00.00.00.00.00.01.00.0
1Super Mario Bros.Platform0.00.00.00.01.00.00.00.00.00.00.00.0
2Mario Kart WiiRacing0.00.00.00.00.00.01.00.00.00.00.00.0
3Wii Sports ResortSports0.00.00.00.00.00.00.00.00.00.01.00.0
4Pokemon Red/Pokemon BlueRole-Playing0.00.00.00.00.00.00.01.00.00.00.00.0
\n", + "
" + ], + "text/plain": [ + " Name Genre Action Adventure Fighting Misc \\\n", + "0 Wii Sports Sports 0.0 0.0 0.0 0.0 \n", + "1 Super Mario Bros. Platform 0.0 0.0 0.0 0.0 \n", + "2 Mario Kart Wii Racing 0.0 0.0 0.0 0.0 \n", + "3 Wii Sports Resort Sports 0.0 0.0 0.0 0.0 \n", + "4 Pokemon Red/Pokemon Blue Role-Playing 0.0 0.0 0.0 0.0 \n", + "\n", + " Platform Puzzle Racing Role-Playing Shooter Simulation Sports \\\n", + "0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 \n", + "1 1.0 0.0 0.0 0.0 0.0 0.0 0.0 \n", + "2 0.0 0.0 1.0 0.0 0.0 0.0 0.0 \n", + "3 0.0 0.0 0.0 0.0 0.0 0.0 1.0 \n", + "4 0.0 0.0 0.0 1.0 0.0 0.0 0.0 \n", + "\n", + " Strategy \n", + "0 0.0 \n", + "1 0.0 \n", + "2 0.0 \n", + "3 0.0 \n", + "4 0.0 " + ] + }, + "execution_count": 11, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "vg_df_ohe = pd.concat([vg_df_2,gen_features],axis=1) # 两个数据合并\n", + "vg_df_ohe.head() # 可以看到Platform列第二行为1，对应着Genre列第二行是Platform字符串" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## Get Dummy\n", + "更加实用的onehot" + ] + }, + { + "cell_type": "code", + "execution_count": 12, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "(16598, 13)\n" + ] + }, + { + "data": { + "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", + "
NameGenreAdventureFightingMiscPlatformPuzzleRacingRole-PlayingShooterSimulationSportsStrategy
0Wii SportsSports00000000010
1Super Mario Bros.Platform00010000000
2Mario Kart WiiRacing00000100000
3Wii Sports ResortSports00000000010
4Pokemon Red/Pokemon BlueRole-Playing00000010000
\n", + "
" + ], + "text/plain": [ + " Name Genre Adventure Fighting Misc \\\n", + "0 Wii Sports Sports 0 0 0 \n", + "1 Super Mario Bros. Platform 0 0 0 \n", + "2 Mario Kart Wii Racing 0 0 0 \n", + "3 Wii Sports Resort Sports 0 0 0 \n", + "4 Pokemon Red/Pokemon Blue Role-Playing 0 0 0 \n", + "\n", + " Platform Puzzle Racing Role-Playing Shooter Simulation Sports \\\n", + "0 0 0 0 0 0 0 1 \n", + "1 1 0 0 0 0 0 0 \n", + "2 0 0 1 0 0 0 0 \n", + "3 0 0 0 0 0 0 1 \n", + "4 0 0 0 1 0 0 0 \n", + "\n", + " Strategy \n", + "0 0 \n", + "1 0 \n", + "2 0 \n", + "3 0 \n", + "4 0 " + ] + }, + "execution_count": 12, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "gen_dummy_features = pd.get_dummies(vg_df['Genre'],drop_first=True) # drop_first=True删掉全为0的列\n", + "dummy_df = pd.concat([vg_df[['Name', 'Genre']], gen_dummy_features], axis=1)\n", + "print(dummy_df.shape)\n", + "dummy_df.head()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "可以看到两句话就解决了我们上面那一长串" + ] + }, + { + "cell_type": "code", + "execution_count": 13, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "(16598, 14)\n" + ] + }, + { + "data": { + "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", + "
NameGenreActionAdventureFightingMiscPlatformPuzzleRacingRole-PlayingShooterSimulationSportsStrategy
0Wii SportsSports000000000010
1Super Mario Bros.Platform000010000000
2Mario Kart WiiRacing000000100000
3Wii Sports ResortSports000000000010
4Pokemon Red/Pokemon BlueRole-Playing000000010000
\n", + "
" + ], + "text/plain": [ + " Name Genre Action Adventure Fighting Misc \\\n", + "0 Wii Sports Sports 0 0 0 0 \n", + "1 Super Mario Bros. Platform 0 0 0 0 \n", + "2 Mario Kart Wii Racing 0 0 0 0 \n", + "3 Wii Sports Resort Sports 0 0 0 0 \n", + "4 Pokemon Red/Pokemon Blue Role-Playing 0 0 0 0 \n", + "\n", + " Platform Puzzle Racing Role-Playing Shooter Simulation Sports \\\n", + "0 0 0 0 0 0 0 1 \n", + "1 1 0 0 0 0 0 0 \n", + "2 0 0 1 0 0 0 0 \n", + "3 0 0 0 0 0 0 1 \n", + "4 0 0 0 1 0 0 0 \n", + "\n", + " Strategy \n", + "0 0 \n", + "1 0 \n", + "2 0 \n", + "3 0 \n", + "4 0 " + ] + }, + "execution_count": 13, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "gen_dummy_features = pd.get_dummies(vg_df['Genre']) # 和上面相比少了drop_first=True，一般用这种\n", + "dummy_df_true = pd.concat([vg_df[['Name', 'Genre']], gen_dummy_features], axis=1)\n", + "print(dummy_df_true.shape)\n", + "dummy_df_true.head()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## 二值特征化" + ] + }, + { + "cell_type": "code", + "execution_count": 14, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
NameYear
0Wii Sports2006.0
1Super Mario Bros.1985.0
2Mario Kart Wii2008.0
3Wii Sports Resort2009.0
4Pokemon Red/Pokemon Blue1996.0
\n", + "
" + ], + "text/plain": [ + " Name Year\n", + "0 Wii Sports 2006.0\n", + "1 Super Mario Bros. 1985.0\n", + "2 Mario Kart Wii 2008.0\n", + "3 Wii Sports Resort 2009.0\n", + "4 Pokemon Red/Pokemon Blue 1996.0" + ] + }, + "execution_count": 14, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "vg_year_df = vg_df[['Name', 'Year']]\n", + "vg_year_df.head()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "我们把2000年以上的归类为1，其它归类为0" + ] + }, + { + "cell_type": "code", + "execution_count": 15, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
NameYearYear_tow
0Wii Sports2006.01
1Super Mario Bros.1985.00
2Mario Kart Wii2008.01
3Wii Sports Resort2009.01
4Pokemon Red/Pokemon Blue1996.00
\n", + "
" + ], + "text/plain": [ + " Name Year Year_tow\n", + "0 Wii Sports 2006.0 1\n", + "1 Super Mario Bros. 1985.0 0\n", + "2 Mario Kart Wii 2008.0 1\n", + "3 Wii Sports Resort 2009.0 1\n", + "4 Pokemon Red/Pokemon Blue 1996.0 0" + ] + }, + "execution_count": 15, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "vg_year_df['Year_tow'] = np.where(vg_year_df['Year'] >= 2000, 1, 0)\n", + "vg_year_df.head()" + ] + }, + { + "cell_type": "code", + "execution_count": 16, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
NameYearYear_towbn_year
0Wii Sports2006.011.0
1Super Mario Bros.1985.000.0
2Mario Kart Wii2008.011.0
3Wii Sports Resort2009.011.0
4Pokemon Red/Pokemon Blue1996.000.0
\n", + "
" + ], + "text/plain": [ + " Name Year Year_tow bn_year\n", + "0 Wii Sports 2006.0 1 1.0\n", + "1 Super Mario Bros. 1985.0 0 0.0\n", + "2 Mario Kart Wii 2008.0 1 1.0\n", + "3 Wii Sports Resort 2009.0 1 1.0\n", + "4 Pokemon Red/Pokemon Blue 1996.0 0 0.0" + ] + }, + "execution_count": 16, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "from sklearn.preprocessing import Binarizer\n", + "# sklearn中的方法\n", + "bn = Binarizer(threshold=2000) # 大于2000我1，小于为0\n", + "vg_year_df['Year']=vg_year_df['Year'].fillna(0) # 数据中有Nan值，需要补0，否则无法二分\n", + "bn_year = bn.transform([vg_year_df['Year']])[0] # 获取转换的值，取第0列\n", + "vg_year_df['bn_year'] = bn_year # 插入数据\n", + "vg_year_df.head() # 结果与手动一致" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## 多项式特征\n", + "获得特征的更高维度和互相间关系的项。" + ] + }, + { + "cell_type": "code", + "execution_count": 17, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
NA_SalesEU_Sales
041.4929.02
129.083.58
215.8512.88
315.7511.01
411.278.89
\n", + "
" + ], + "text/plain": [ + " NA_Sales EU_Sales\n", + "0 41.49 29.02\n", + "1 29.08 3.58\n", + "2 15.85 12.88\n", + "3 15.75 11.01\n", + "4 11.27 8.89" + ] + }, + "execution_count": 17, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "polynomial_df = vg_df[['NA_Sales', 'EU_Sales']]\n", + "polynomial_df.head()" + ] + }, + { + "cell_type": "code", + "execution_count": 18, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "array([[4.1490000e+01, 2.9020000e+01, 1.7214201e+03, 1.2040398e+03,\n", + " 8.4216040e+02],\n", + " [2.9080000e+01, 3.5800000e+00, 8.4564640e+02, 1.0410640e+02,\n", + " 1.2816400e+01],\n", + " [1.5850000e+01, 1.2880000e+01, 2.5122250e+02, 2.0414800e+02,\n", + " 1.6589440e+02],\n", + " ...,\n", + " [0.0000000e+00, 0.0000000e+00, 0.0000000e+00, 0.0000000e+00,\n", + " 0.0000000e+00],\n", + " [0.0000000e+00, 1.0000000e-02, 0.0000000e+00, 0.0000000e+00,\n", + " 1.0000000e-04],\n", + " [1.0000000e-02, 0.0000000e+00, 1.0000000e-04, 0.0000000e+00,\n", + " 0.0000000e+00]])" + ] + }, + "execution_count": 18, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "from sklearn.preprocessing import PolynomialFeatures\n", + "\n", + "# degree二次幂的复杂度\n", + "pf = PolynomialFeatures(degree=2, interaction_only=False, include_bias=False)\n", + "res = pf.fit_transform(polynomial_df)\n", + "res" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "以第一行为例：\n", + "
第一列和第二列分别表示原先的第一列和第二列\n", + "
第三列和第五列表示第一列和第二列分别的平方，第四列表示两者的乘积" + ] + }, + { + "cell_type": "code", + "execution_count": 19, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
NA_SalesEU_SalesNA_Sales^2NA_Sales*EU_SalesEU_Sales^2
041.4929.021721.42011204.0398842.1604
129.083.58845.6464104.106412.8164
215.8512.88251.2225204.1480165.8944
315.7511.01248.0625173.4075121.2201
411.278.89127.0129100.190379.0321
\n", + "
" + ], + "text/plain": [ + " NA_Sales EU_Sales NA_Sales^2 NA_Sales*EU_Sales EU_Sales^2\n", + "0 41.49 29.02 1721.4201 1204.0398 842.1604\n", + "1 29.08 3.58 845.6464 104.1064 12.8164\n", + "2 15.85 12.88 251.2225 204.1480 165.8944\n", + "3 15.75 11.01 248.0625 173.4075 121.2201\n", + "4 11.27 8.89 127.0129 100.1903 79.0321" + ] + }, + "execution_count": 19, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "intr_features = pd.DataFrame(res, columns=['NA_Sales',\n", + " 'EU_Sales',\n", + " 'NA_Sales^2',\n", + " 'NA_Sales*EU_Sales',\n", + " 'EU_Sales^2'])\n", + "intr_features.head()" + ] + }, + { + "cell_type": "code", + "execution_count": 20, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
RankNamePlatformYearGenrePublisherNA_SalesEU_SalesJP_SalesOther_SalesGlobal_SalesGenreLabelGenreMap
01Wii SportsWii2006.0SportsNintendo41.4929.023.778.4682.741010
12Super Mario Bros.NES1985.0PlatformNintendo29.083.586.810.7740.2444
23Mario Kart WiiWii2008.0RacingNintendo15.8512.883.793.3135.8266
34Wii Sports ResortWii2009.0SportsNintendo15.7511.013.282.9633.001010
45Pokemon Red/Pokemon BlueGB1996.0Role-PlayingNintendo11.278.8910.221.0031.3777
\n", + "
" + ], + "text/plain": [ + " Rank Name Platform Year Genre Publisher \\\n", + "0 1 Wii Sports Wii 2006.0 Sports Nintendo \n", + "1 2 Super Mario Bros. NES 1985.0 Platform Nintendo \n", + "2 3 Mario Kart Wii Wii 2008.0 Racing Nintendo \n", + "3 4 Wii Sports Resort Wii 2009.0 Sports Nintendo \n", + "4 5 Pokemon Red/Pokemon Blue GB 1996.0 Role-Playing Nintendo \n", + "\n", + " NA_Sales EU_Sales JP_Sales Other_Sales Global_Sales GenreLabel \\\n", + "0 41.49 29.02 3.77 8.46 82.74 10 \n", + "1 29.08 3.58 6.81 0.77 40.24 4 \n", + "2 15.85 12.88 3.79 3.31 35.82 6 \n", + "3 15.75 11.01 3.28 2.96 33.00 10 \n", + "4 11.27 8.89 10.22 1.00 31.37 7 \n", + "\n", + " GenreMap \n", + "0 10 \n", + "1 4 \n", + "2 6 \n", + "3 10 \n", + "4 7 " + ] + }, + "execution_count": 20, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "vg_df.head()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## Binning 特征\n", + "一般用来处理年龄" + ] + }, + { + "cell_type": "code", + "execution_count": 21, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
NameYear
0Wii Sports2006.0
1Super Mario Bros.1985.0
2Mario Kart Wii2008.0
3Wii Sports Resort2009.0
4Pokemon Red/Pokemon Blue1996.0
\n", + "
" + ], + "text/plain": [ + " Name Year\n", + "0 Wii Sports 2006.0\n", + "1 Super Mario Bros. 1985.0\n", + "2 Mario Kart Wii 2008.0\n", + "3 Wii Sports Resort 2009.0\n", + "4 Pokemon Red/Pokemon Blue 1996.0" + ] + }, + "execution_count": 21, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "bin_df = vg_df[['Name','Year']] # 假设GenreLabel是年龄\n", + "bin_df.head()" + ] + }, + { + "cell_type": "code", + "execution_count": 24, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "Text(0, 0.5, 'Frequency')" + ] + }, + "execution_count": 24, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "import matplotlib.pyplot as plt\n", + "import matplotlib as mpl\n", + "import scipy.stats as spstats\n", + "\n", + "fig, ax = plt.subplots()\n", + "bin_df['Year'].hist(color='#A9C5D3')\n", + "ax.set_title('Developer Global_Sales Hostogram', fontsize=12)\n", + "ax.set_xlabel('Global_Sales', fontsize=12)\n", + "ax.set_ylabel('Frequency', fontsize=12)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "这样区间就出来了，我们可以分成多个区间，如1980-1985是一个区间，1986-1990是一个区间" + ] + }, + { + "cell_type": "code", + "execution_count": 23, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
NameYearYear_bin
0Wii Sports2006.05
1Super Mario Bros.1985.01
2Mario Kart Wii2008.06
3Wii Sports Resort2009.06
4Pokemon Red/Pokemon Blue1996.03
5Tetris1989.02
6New Super Mario Bros.2006.05
7Wii Play2006.05
8New Super Mario Bros. Wii2009.06
9Duck Hunt1984.00
\n", + "
" + ], + "text/plain": [ + " Name Year Year_bin\n", + "0 Wii Sports 2006.0 5\n", + "1 Super Mario Bros. 1985.0 1\n", + "2 Mario Kart Wii 2008.0 6\n", + "3 Wii Sports Resort 2009.0 6\n", + "4 Pokemon Red/Pokemon Blue 1996.0 3\n", + "5 Tetris 1989.0 2\n", + "6 New Super Mario Bros. 2006.0 5\n", + "7 Wii Play 2006.0 5\n", + "8 New Super Mario Bros. Wii 2009.0 6\n", + "9 Duck Hunt 1984.0 0" + ] + }, + "execution_count": 23, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "gle = LabelEncoder() # 实例化\n", + "bin_df['Year_bin'] = pd.cut(bin_df['Year'], 9) # 切分成9组，也可以自己指定切分区间\n", + "bin_df['Year_bin'] = bin_df['Year_bin'].astype(str) # 转换类型为字符串\n", + "bin_year = gle.fit_transform(bin_df['Year_bin']) # 利用LabelEncoder方法变成1-9的数值\n", + "bin_df['Year_bin'] = bin_year # 赋值到新的列\n", + "bin_df.head(10)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## 对数变换\n", + "\n", + "经常有这样的假设：数据的分布是正态分布。如线性回归的时候误差项要满足正态分布，而当数据不满足的时候，则需要把数据变换成正态分布" + ] + }, + { + "cell_type": "code", + "execution_count": 37, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
NameNA_SalesNA_Sales_log
0Wii Sports41.493.749269
1Super Mario Bros.29.083.403860
2Mario Kart Wii15.852.824351
3Wii Sports Resort15.752.818398
4Pokemon Red/Pokemon Blue11.272.507157
\n", + "
" + ], + "text/plain": [ + " Name NA_Sales NA_Sales_log\n", + "0 Wii Sports 41.49 3.749269\n", + "1 Super Mario Bros. 29.08 3.403860\n", + "2 Mario Kart Wii 15.85 2.824351\n", + "3 Wii Sports Resort 15.75 2.818398\n", + "4 Pokemon Red/Pokemon Blue 11.27 2.507157" + ] + }, + "execution_count": 37, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "df_log = vg_df[['Name','NA_Sales']] \n", + "df_log['NA_Sales_log'] = np.log((1+df_log['NA_Sales']))\n", + "df_log.head()" + ] + }, + { + "cell_type": "code", + "execution_count": 45, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "" + ] + }, + "execution_count": 45, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "# 画两张对比图，左边log过的更偏态\n", + "fig, ax = plt.subplots()\n", + "plt.subplot(121) \n", + "df_log['NA_Sales_log'].hist(color='#A9C5D3')\n", + "\n", + "plt.subplot(122) \n", + "df_log['NA_Sales'].hist(color='#A9C5D3')" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "上面是手动的，还有模块化的BoxCox，这里暂不做示例" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## 日期相关特征\n", + "将时间特征转换成可以应用的数据" + ] + }, + { + "cell_type": "code", + "execution_count": 46, + "metadata": {}, + "outputs": [], + "source": [ + "import datetime\n", + "from dateutil.parser import parse\n", + "import pytz" + ] + }, + { + "cell_type": "code", + "execution_count": 55, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
Time
02020-12-16 10:30:00.360000+00:00
12019-04-16 12:15:00.250000+00:00
22018-10-16 08:30:00.750000+00:00
32019-01-16 23:30:00.255500+00:00
\n", + "
" + ], + "text/plain": [ + " Time\n", + "0 2020-12-16 10:30:00.360000+00:00\n", + "1 2019-04-16 12:15:00.250000+00:00\n", + "2 2018-10-16 08:30:00.750000+00:00\n", + "3 2019-01-16 23:30:00.255500+00:00" + ] + }, + "execution_count": 55, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "time_stamps = ['2020-12-16 10:30:00.360000+00:00','2019-04-16 12:15:00.250000+00:00',\n", + " '2018-10-16 08:30:00.750000+00:00','2019-01-16 23:30:00.255500+00:00']\n", + "\n", + "df = pd.DataFrame(time_stamps, columns=['Time'])\n", + "df" + ] + }, + { + "cell_type": "code", + "execution_count": 56, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "array([Timestamp('2020-12-16 10:30:00.360000+0000', tz='UTC'),\n", + " Timestamp('2019-04-16 12:15:00.250000+0000', tz='UTC'),\n", + " Timestamp('2018-10-16 08:30:00.750000+0000', tz='UTC'),\n", + " Timestamp('2019-01-16 23:30:00.255500+0000', tz='UTC')],\n", + " dtype=object)" + ] + }, + "execution_count": 56, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "ts_objs = np.array([pd.Timestamp(item) for item in np.array(df.Time)])\n", + "df['TS_obj'] = ts_objs\n", + "ts_objs" + ] + }, + { + "cell_type": "code", + "execution_count": 59, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
TimeYearMonthdayDayOfWeekWeekDayNameDayOfYearWeekOfYearQuarter
02020-12-16 10:30:00.360000+00:00202012162Wednesday351514
12019-04-16 12:15:00.250000+00:0020194161Tuesday106162
22018-10-16 08:30:00.750000+00:00201810161Tuesday289424
32019-01-16 23:30:00.255500+00:0020191162Wednesday1631
\n", + "
" + ], + "text/plain": [ + " Time Year Month day DayOfWeek WeekDayName \\\n", + "0 2020-12-16 10:30:00.360000+00:00 2020 12 16 2 Wednesday \n", + "1 2019-04-16 12:15:00.250000+00:00 2019 4 16 1 Tuesday \n", + "2 2018-10-16 08:30:00.750000+00:00 2018 10 16 1 Tuesday \n", + "3 2019-01-16 23:30:00.255500+00:00 2019 1 16 2 Wednesday \n", + "\n", + " DayOfYear WeekOfYear Quarter \n", + "0 351 51 4 \n", + "1 106 16 2 \n", + "2 289 42 4 \n", + "3 16 3 1 " + ] + }, + "execution_count": 59, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "df['Year'] = df['TS_obj'].apply(lambda d: d.year)\n", + "df['Month'] = df['TS_obj'].apply(lambda d: d.month)\n", + "df['Day'] = df['TS_obj'].apply(lambda d: d.day)\n", + "df['DayOfWeek'] = df['TS_obj'].apply(lambda d: d.dayofweek)\n", + "df['WeekDayName'] = df['TS_obj'].apply(lambda d: d.weekday_name)\n", + "df['DayOfYear'] = df['TS_obj'].apply(lambda d: d.dayofyear)\n", + "df['WeekOfYear'] = df['TS_obj'].apply(lambda d: d.weekofyear)\n", + "df['Quarter'] = df['TS_obj'].apply(lambda d: d.quarter)\n", + "\n", + "df[['Time','Year','Month','day','DayOfWeek','WeekDayName','DayOfYear','WeekOfYear','Quarter']]" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "这样就能从时间数据中获取很多数据，不同场景对不同数据有需求，如外卖则会关注周末和季节等。" + ] + }, + { + "cell_type": "code", + "execution_count": 60, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
TimeHourMinuteSecondMicrosecondUtcoffset
02020-12-16 10:30:00.360000+00:00103003600000 days
12019-04-16 12:15:00.250000+00:00121502500000 days
22018-10-16 08:30:00.750000+00:0083007500000 days
32019-01-16 23:30:00.255500+00:00233002555000 days
\n", + "
" + ], + "text/plain": [ + " Time Hour Minute Second Microsecond \\\n", + "0 2020-12-16 10:30:00.360000+00:00 10 30 0 360000 \n", + "1 2019-04-16 12:15:00.250000+00:00 12 15 0 250000 \n", + "2 2018-10-16 08:30:00.750000+00:00 8 30 0 750000 \n", + "3 2019-01-16 23:30:00.255500+00:00 23 30 0 255500 \n", + "\n", + " Utcoffset \n", + "0 0 days \n", + "1 0 days \n", + "2 0 days \n", + "3 0 days " + ] + }, + "execution_count": 60, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "df['Hour'] = df['TS_obj'].apply(lambda d: d.hour)\n", + "df['Minute'] = df['TS_obj'].apply(lambda d: d.minute)\n", + "df['Second'] = df['TS_obj'].apply(lambda d: d.second)\n", + "df['Microsecond'] = df['TS_obj'].apply(lambda d: d.microsecond)\n", + "df['Utcoffset'] = df['TS_obj'].apply(lambda d: d.utcoffset()) # UTC时间位移\n", + "\n", + "df[['Time','Hour','Minute','Second','Microsecond','Utcoffset']]" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "又比如按早晚切分时间" + ] + }, + { + "cell_type": "code", + "execution_count": 61, + "metadata": {}, + "outputs": [ + { + "data": { + "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", + "
TimeHourTimeOfDayBin
02020-12-16 10:30:00.360000+00:0010Morning
12019-04-16 12:15:00.250000+00:0012Afternoon
22018-10-16 08:30:00.750000+00:008Morning
32019-01-16 23:30:00.255500+00:0023Night
\n", + "
" + ], + "text/plain": [ + " Time Hour TimeOfDayBin\n", + "0 2020-12-16 10:30:00.360000+00:00 10 Morning\n", + "1 2019-04-16 12:15:00.250000+00:00 12 Afternoon\n", + "2 2018-10-16 08:30:00.750000+00:00 8 Morning\n", + "3 2019-01-16 23:30:00.255500+00:00 23 Night" + ] + }, + "execution_count": 61, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "hour_bins = [-1, 5, 11, 16, 21, 23]\n", + "bin_names = ['Late Night', 'Morning', 'Afternoon', 'Evening', 'Night']\n", + "df['TimeOfDayBin'] = pd.cut(df['Hour'],bins=hour_bins,labels=bin_names)\n", + "\n", + "df[['Time','Hour','TimeOfDayBin']]" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3", + "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": "3.7.3" + } + }, + "nbformat": 4, + "nbformat_minor": 2 +}