> For the complete documentation index, see [llms.txt](https://jen-hsuan-hsieh.gitbook.io/python/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://jen-hsuan-hsieh.gitbook.io/python/chapter-2courses/21python-for-data-science-and-machine-learning-bootcamp/qwqw/21122logistic-regression-with-python.md). # 2.1.12.2.Logistic regression with Python ## 1. Titanic: Machine Learning from Disaster * 在[資料頁面](https://www.kaggle.com/c/titanic/data)按下Download, 即可下載CSV檔 ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-23%20下午7.22.52.png) * 從頁面上可得到資料欄位的說明 ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-23%20下午7.49.01.png) ## 2.資料探勘及分析 1. 匯入基本的library * [pandas](https://jenhsuan.gitbooks.io/python/content/chapter-2courses/21python-for-data-science-and-machine-learning-bootcamp/211jupyter-overview/214python-for-data-analysis-pandas.html), [numpy](https://jenhsuan.gitbooks.io/python/content/chapter-2courses/21python-for-data-science-and-machine-learning-bootcamp/211jupyter-overview/213python-for-data-analysis-numpy.html), [matplotlib](https://jenhsuan.gitbooks.io/python/content/test.html), [seaborn](https://jenhsuan.gitbooks.io/python/content/217python-for-data-visualization-seaborn.html) ``` import pandas as pd import numpy as np import matplotlib.pyplot as plt import seaborn as sns ``` 2. 將圖表直接嵌入到Notebook之中 ``` %matplotlib inline ``` 3. 讀取資料並了解資料 * 讀取資料 ``` train = pd.read_csv('USA_Housing.csv') ``` * 先了解資料欄位的型別以及變數的型態, 由pd.info()可以知道這份資料有9個欄位: 有2筆屬於float, 4筆屬於int64, 3筆屬於uint8 ``` train.info() ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-23%20下午7.57.26.png) * 可以確認前幾筆資料的內容長怎樣, 例如確認前10筆 ``` train.head(10) ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-23%20下午7.59.31.png) 4. 視覺化分析 * [Seaborn Heatmap: 創建一個映射圖以了解缺少的資料](https://jenhsuan.gitbooks.io/python/content/217python-for-data-visualization-seaborn/2174matrix-plots.html) * 可以發現Age, Cabin有一些屬於null的資料 ``` sns.heatmap(train.isnull(), yticklabels = False, cbar = False, cmap='viridis') ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-23%20下午8.08.15.png) * [Seaborn countplot: 瞭解倖存者的比例](https://jenhsuan.gitbooks.io/python/content/217python-for-data-visualization-seaborn/2173categorical-plots.html) * 表示生還者 ``` sns.set_style('whitegrid') sns.countplot(x = 'Survived', data = train) ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-23%20下午8.14.41.png) * [Seaborn countplot:瞭解倖存者/非倖存者的男女比例](https://jenhsuan.gitbooks.io/python/content/217python-for-data-visualization-seaborn/2173categorical-plots.html) ``` sns.countplot(x = 'Survived', hue = 'Sex', data = train, palette='RdBu_r') ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-23%20下午9.31.04.png) * [Seaborn countplot:瞭解倖存者/非倖存者的上船港口比例](https://jenhsuan.gitbooks.io/python/content/217python-for-data-visualization-seaborn/2173categorical-plots.html) ``` sns.countplot(x = 'Survived', hue = 'Embarked', data = train, palette='RdBu_r') ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-23%20下午9.49.12.png) * [Seaborn countplot:瞭解倖存者/非倖存者跟艙等間的關係](https://jenhsuan.gitbooks.io/python/content/217python-for-data-visualization-seaborn/2173categorical-plots.html) ``` sns.countplot(x = 'Survived', hue = 'Pclass', data = train, palette='RdBu_r') ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-23%20下午9.53.43.png) * [Seaborn distplot: 瞭解船上的乘客年齡分佈](https://jenhsuan.gitbooks.io/python/content/217python-for-data-visualization-seaborn/2172distribution-plot.html) ``` sns.distplot(train['Age'].dropna(), kde = False, bins = 30) ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-23%20下午9.57.09.png) * [Seaborn countplot:船上的人是否有兄弟姊妹: 大多數的人都是單身](https://jenhsuan.gitbooks.io/python/content/217python-for-data-visualization-seaborn/2173categorical-plots.html) ``` sns.countplot(x = 'SibSp', data = train) ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-23%20下午10.03.01.png) * [pandas plot: 乘客票價區間的分佈](https://jenhsuan.gitbooks.io/python/content/2323.html) ``` train['Fare'].plot.hist(bins=40, figsize = (10, 4)) ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-23%20下午10.26.33.png) * [Plotly and Cufflinks: 建立互動式的圖表](https://jenhsuan.gitbooks.io/python/content/wqew/2182plotly-and-cufflinks.html) ``` import cufflinks as cf cf.go_offline() train['Fare'].iplot(bins=40) ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-23%20下午10.17.39.png) * [Seaborn boxplot: 可以看出不同等級的機艙乘客的年齡分佈](https://jenhsuan.gitbooks.io/python/content/217python-for-data-visualization-seaborn/2173categorical-plots.html) ``` plt.figure(figsize=(10,7)) sns.boxplot(x='Pclass', y='Age',data=train) ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-23%20下午10.37.45.png) ## 3.分類及處理不存在的資料 * 在進行訓練之前, 需要先檢視原始資料是否有缺漏, 並設法補充缺漏的地方 * 從heat map中可以發現Age, Cabin有一些屬於null的資料 ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-23%20下午8.08.15.png) * 先對Age資料做處理: 目的是填充NA的資料 * 定義一個function用來替換掉null值的資料 ``` #將null值換成假資料 def impute_age(cols): Age = cols[0] Pclass = cols[1] if pd.isnull(Age): if Pclass == 1: return 37 elif Pclass == 2: return 29 else: return 24 else: return Age train['Age'] = train[['Age', 'Pclass']].apply(impute_age, axis = 1) #重新畫heat map sns.heatmap(train.isnull(), yticklabels = False, cbar = False, cmap='viridis') ``` * 重新畫出的heat map可以看出, Age的部分的na已經處理好了 ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-24%20上午8.21.22.png) * 接著處理Cabin: 目的是丟棄NA的資料 * drop Cabin ``` train.drop('Cabin', axis = 1, inplace = True) #drop後仍有一些NA sns.heatmap(train.isnull(), yticklabels = False, cbar = False, cmap='viridis') ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-24%20上午8.32.15.png) * drop NA的column ``` train.dropna(inplace = True) sns.heatmap(train.isnull(), yticklabels = False, cbar = False, cmap='viridis') ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-24%20上午8.34.04.png) * 對於某些離散特徵, 例如Sex, Embarked, 可以[轉成one-hot編碼](https://blog.csdn.net/lujiandong1/article/details/52836051), 目的是利用one-hot編碼數值化這些離散資料以進行後面的training * 處理Sex資料 ``` #性別 #one hot 編碼 sex = pd.get_dummies(train['Sex'], drop_first=True) sex.head() ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-24%20上午8.42.51.png) * 處理Embarked資料 ``` #登船港口 embark = pd.get_dummies(train['Embarked'], drop_first=True) embark.head() ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-24%20上午8.44.14.png) * 將原始資料跟one-hot的資料concat在一起 ``` train = pd.concat([train, sex, embark], axis = 1) train.head() ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-24%20上午8.49.33.png) * 從資料中移除原始的Sex, Embarked資料, 此外移除不需要的Name, Ticket, PassengerId ``` train.drop(['Sex', 'Embarked', 'Name', 'Ticket', 'PassengerId'], axis = 1, inplace = True) train.head() ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-24%20上午8.50.27.png) ## 4.訓練及建模以預測乘客是否會生還 * 接著用除了Survived以外的資料來預測Survived, 在實際案例時可以使用另一批資料做為測試資料. 這邊使用的是[sklearn](https://jenhsuan.gitbooks.io/python/content/chapter-2courses/21python-for-data-science-and-machine-learning-bootcamp/211jupyter-overview/2111linear-regression-with-python/linear-regression-with-python.html) * 首先用train\_test\_split, 這個函式可以隨機劃分訓練集和測試集 ``` X = train.drop('Survived', axis = 1) Y = train['Survived'] from sklearn.cross_validation import train_test_split X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=0.4, random_state=101) ``` * 接著使用Logic regression的library進行訓練, 接著進行預測 ``` from sklearn.linear_model import LogisticRegression logmodel = LogisticRegression() logmodel.fit(X_train, y_train) predictions = logmodel.predict(X_test) ``` * Evaluate model ``` from sklearn.metrics import classification_report print(classification_report(y_test, predictions)) ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-24%20上午9.08.52.png) * [confusion matrix](https://jenhsuan.gitbooks.io/python/content/qwqw/21121logistic-regression-theory.html) ``` from sklearn.metrics import confusion_matrix confusion_matrix(y_test, predictions) ``` ![](https://github.com/jenhsuan/python/tree/8fc9c0b8df4ccd709d3078c2d8842af0932de09d/assets/螢幕快照%202018-06-24%20上午9.09.58.png)