美国人口普查年收入比赛程序

# 本程序参考网页： # https://blog.csdn.net/randompeople/article/details/81429264 # 数据已包含在同名文件夹中，且在上述网页中可以找到下载链接。 # 王磊-中国石油大学（北京） 2019-3-4 18:19:16 # coding: utf-8 # In[9]: import pandas as pd # load csv's (pd.read_csv) import numpy as np # math (lin. algebra) import sklearn as skl # machine learning from sklearn.ensemble import RandomForestClassifier #from plotnine import * import matplotlib.pyplot as plt from sklearn.preprocessing import LabelEncoder from sklearn_pandas import DataFrameMapper from sklearn.tree import DecisionTreeClassifier from sklearn.metrics import classification_report # In[10]: def get_train_data(): #下面两个文件路径替换为你电脑上该文件的路径 train_path = "D:/workspace/Data/kggal/AmericaIncome/adult.data" test_path = 'D:/workspace/Data/kggal/AmericaIncome/adult.test' columns = ['Age','Workclass','fnlgwt','Education','EdNum','MaritalStatus','Occupation','Relationship','Race','Sex','CapitalGain','CapitalLoss','HoursPerWeek','Country','Income'] df_train_set = pd.read_csv(train_path, names=columns) #print(df_train_set.head()) return df_train_set def get_test_data(): test_path = 'D:/workspace/Data/kggal/AmericaIncome/adult.test' columns = ['Age','Workclass','fnlgwt','Education','EdNum','MaritalStatus','Occupation','Relationship','Race','Sex','CapitalGain','CapitalLoss','HoursPerWeek','Country','Income'] df_test_set = pd.read_csv(test_path, names=columns) #print(df_test_set.head()) return df_test_set # In[13]: df_train_set=get_train_data() df_test_set=get_test_data() df_train_set.drop('fnlgwt', axis=1, inplace=True) df_test_set.drop('fnlgwt', axis=1, inplace=True) print(df_train_set.replace(' ?', np.nan).shape) #(32561, 14) print(df_train_set.replace(' ?', np.nan).dropna().shape) # (30162, 15) print(df_test_set.replace(' ?', np.nan).shape) #(16282, 14) print(df_test_set.replace(' ?', np.nan).dropna().shape) # (15060, 15) # 删除含有?（缺失行） train_set = df_train_set.replace(' ?', np.nan).dropna() test_set = df_test_set.replace(' ?', np.nan).dropna() # 把测试语料集中的 Income 归一化 test_set['Income'] = test_set.Income.replace({' <=50K.': ' <=50K', ' >50K.': ' >50K'}) print(test_set.Income.unique()) # [' <=50K' ' >50K'] print(df_train_set.Income.unique()) # [' <=50K' ' >50K'] # 因为有 受教育的年数，所以这里不需要教育这一列 train_set.drop(["Education"], axis=1, inplace=True) test_set.drop(["Education"], axis=1, inplace=True) # In[14]: combined_set = pd.concat([train_set, test_set], axis=0) for feature in combined_set.columns: if combined_set[feature].dtype == 'object': combined_set[feature] = pd.Categorical(combined_set[feature]).codes train_set = combined_set[:train_set.shape[0]] test_set = combined_set[test_set.shape[0]:] print(train_set.Workclass.unique()) print(test_set.Income.unique()) cols = list(train_set.columns) cols.remove("Income") x_train, y_train = train_set[cols].values, train_set["Income"].values #测试集的输入和输出结果分开，用来校验模型的准确率 x_test, y_test = test_set[cols].values, test_set["Income"].values # In[15]: treeClassifier = DecisionTreeClassifier() treeClassifier.fit(x_train, y_train) # 训练模型 treeClassifier.score(x_test, y_test) print(treeClassifier.score(x_test, y_test)) # 输出为：0.8700351435581195 y_pred = treeClassifier.predict(x_test) # 用测试集做预测 print(classification_report(y_test, y_pred)) # 查看模型的预测值与真实值 #输出为 # In[ ]: