- \n
- \n
\u6570\u636e\u5904\u7406：Python\u3001Pandas（\u6570\u636e\u8bfb\u53d6\u3001\u7a7a\u503c\u5904\u7406\u3001\u683c\u5f0f\u8f6c\u6362）<\/p>\n<\/li>\n
- \n
\u4f20\u7edf\u673a\u5668\u5b66\u4e60：Scikit-learn（\u903b\u8f91\u56de\u5f52\u3001SVM\u3001\u968f\u673a\u68ee\u6797\u3001AdaBoost）<\/p>\n<\/li>\n
- \n
\u96c6\u6210\u5b66\u4e60：XGBoost（\u9ad8\u6027\u80fd\u68af\u5ea6\u63d0\u5347\u6811）<\/p>\n<\/li>\n
- \n
\u6df1\u5ea6\u5b66\u4e60：PyTorch（MLP\u795e\u7ecf\u7f51\u7edc\u30011D-CNN）<\/p>\n<\/li>\n
- \n
\u8bc4\u4f30\u5de5\u5177：Scikit-learn metrics（\u51c6\u786e\u7387\u3001\u7cbe\u786e\u7387\u3001\u53ec\u56de\u7387\u3001F1\u5206\u6570）<\/p>\n<\/li>\n<\/ul>\n
1.3 \u6570\u636e\u96c6\u8bf4\u660e<\/h4>\n
\u6570\u636e\u96c6\u5305\u542b\u77ff\u7269\u7684\u591a\u7ef4\u5ea6\u6210\u5206\u7279\u5f81（\u5982\u5143\u7d20\u542b\u91cf\u3001\u5bc6\u5ea6\u3001\u786c\u5ea6\u7b49），\u6807\u7b7e\u4e3a\u201c\u77ff\u7269\u7c7b\u578b\u201d\u3002\u539f\u59cb\u6570\u636e\u5b58\u5728\u90e8\u5206\u7a7a\u503c，\u9700\u5148\u8fdb\u884c\u9884\u5904\u7406；\u540e\u7eed\u5c06\u6570\u636e\u96c6\u6309\u5e38\u89c4\u6bd4\u4f8b\u5212\u5206\u4e3a\u8bad\u7ec3\u96c6\u4e0e\u6d4b\u8bd5\u96c6，\u7528\u4e8e\u6a21\u578b\u8bad\u7ec3\u4e0e\u8bc4\u4f30\u3002<\/p>\n
![\"\"](\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2026\/01\/20260116165710-696a6de656a8e.png\")
<\/p>\n\u4e8c\u3001\u6838\u5fc3\u6d41\u7a0b：\u4ece\u6570\u636e\u9884\u5904\u7406\u5230\u6a21\u578b\u8bad\u7ec3<\/h3>\n
\u6574\u4e2a\u7cfb\u7edf\u642d\u5efa\u5206\u4e3a\u4e09\u5927\u9636\u6bb5：\u6570\u636e\u9884\u5904\u7406（\u7a7a\u503c\u586b\u5145）\u2192 \u591a\u6a21\u578b\u5206\u7c7b\u8bad\u7ec3 \u2192 \u7ed3\u679c\u5bf9\u6bd4\u4e0e\u6700\u4f18\u65b9\u6848\u7b5b\u9009，\u5171\u8986\u76d66\u79cd\u586b\u5145\u65b9\u6cd5\u00d77\u79cd\u5206\u7c7b\u6a21\u578b，\u5f62\u621042\u79cd\u6280\u672f\u65b9\u6848\u3002<\/p>\n
\u9636\u6bb5\u4e00：\u6570\u636e\u9884\u5904\u7406\u2014\u2014\u7a7a\u503c\u586b\u5145\u7b56\u7565\u5bf9\u6bd4<\/h4>\n
\u539f\u59cb\u6570\u636e\u4e2d\u7684\u7a7a\u503c\u4f1a\u5bfc\u81f4\u6a21\u578b\u8bad\u7ec3\u62a5\u9519\u3001\u7cbe\u5ea6\u4e0b\u964d，\u56e0\u6b64\u9700\u9488\u5bf9\u6027\u9009\u62e9\u586b\u5145\u65b9\u6cd5\u3002\u672c\u6587\u9009\u53d66\u79cd\u4e3b\u6d41\u7a7a\u503c\u586b\u5145\u7b56\u7565，\u517c\u987e\u7b80\u5355\u6027\u4e0e\u9488\u5bf9\u6027：<\/p>\n
- \n
\u4e2d\u4f4d\u6570\u586b\u5145：\u5bf9\u6570\u503c\u578b\u7279\u5f81，\u7528\u4e2d\u4f4d\u6570\u586b\u5145\u7a7a\u503c，\u6297\u5f02\u5e38\u503c\u80fd\u529b\u5f3a，\u9002\u5408\u77ff\u7269\u6210\u5206\u8fd9\u7c7b\u53ef\u80fd\u5b58\u5728\u6781\u7aef\u503c\u7684\u6570\u636e；<\/p>\n<\/li>\n
- \n
\u5747\u503c\u586b\u5145：\u7528\u7279\u5f81\u5747\u503c\u586b\u5145，\u8ba1\u7b97\u7b80\u5355，\u4f46\u6613\u53d7\u5f02\u5e38\u503c\u5f71\u54cd；<\/p>\n<\/li>\n
- \n
\u4f17\u6570\u586b\u5145：\u5bf9\u79bb\u6563\u7279\u5f81\u6216\u5206\u5e03\u96c6\u4e2d\u7684\u6570\u503c\u7279\u5f81，\u7528\u51fa\u73b0\u9891\u7387\u6700\u9ad8\u7684\u503c\u586b\u5145；<\/p>\n<\/li>\n
- \n
\u5220\u9664\u7a7a\u6570\u636e\u884c：\u76f4\u63a5\u5220\u9664\u542b\u7a7a\u503c\u7684\u6837\u672c，\u64cd\u4f5c\u7b80\u5355\u4f46\u4f1a\u4e22\u5931\u6570\u636e，\u4ec5\u9002\u5408\u7a7a\u503c\u5360\u6bd4\u6781\u4f4e\u7684\u573a\u666f；<\/p>\n<\/li>\n
- \n
\u7ebf\u6027\u56de\u5f52\u586b\u5145：\u4ee5\u5176\u4ed6\u7279\u5f81\u4e3a\u8f93\u5165，\u6784\u5efa\u7ebf\u6027\u56de\u5f52\u6a21\u578b\u9884\u6d4b\u7a7a\u503c，\u5229\u7528\u7279\u5f81\u95f4\u7684\u5173\u8054\u5173\u7cfb，\u586b\u5145\u7cbe\u5ea6\u8f83\u9ad8；<\/p>\n<\/li>\n
- \n
\u968f\u673a\u68ee\u6797\u586b\u5145：\u7528\u968f\u673a\u68ee\u6797\u6a21\u578b\u9884\u6d4b\u7a7a\u503c，\u80fd\u6355\u6349\u7279\u5f81\u95f4\u7684\u975e\u7ebf\u6027\u5173\u7cfb，\u9002\u914d\u590d\u6742\u6570\u636e\u5206\u5e03\u3002<\/p>\n<\/li>\n
\u5173\u952e\u5b9e\u73b0：\u9488\u5bf9\u7ebf\u6027\u56de\u5f52\u4e0e\u968f\u673a\u68ee\u6797\u586b\u5145，\u91c7\u7528\u201c\u9010\u5217\u586b\u5145\u201d\u903b\u8f91\u2014\u2014\u6309\u7a7a\u503c\u6570\u91cf\u4ece\u5c0f\u5230\u5927\u6392\u5e8f，\u7528\u5df2\u586b\u5145\u597d\u7684\u7279\u5f81\u9884\u6d4b\u5f53\u524d\u5217\u7a7a\u503c，\u907f\u514d\u6570\u636e\u6cc4\u9732；\u540c\u65f6\u4e25\u683c\u533a\u5206\u8bad\u7ec3\u96c6\u4e0e\u6d4b\u8bd5\u96c6，\u6d4b\u8bd5\u96c6\u7a7a\u503c\u7528\u8bad\u7ec3\u96c6\u8bad\u7ec3\u7684\u6a21\u578b\u9884\u6d4b，\u4fdd\u8bc1\u8bc4\u4f30\u5ba2\u89c2\u6027\u3002<\/p>\n
\u4ee3\u7801：<\/p>\n
import pandas as pd
\nimport numpy as np
\nfrom sklearn.model_selection import train_test_split
\nimport fill
\nfrom sklearn.preprocessing import StandardScaler
\nfrom imblearn.over_sampling import SMOTE<\/p>\n# 1. \u8bfb\u53d6Excel\u6587\u4ef6
\ndf = pd.read_excel("\u77ff\u7269\u6570\u636e.xls")<\/p>\n# 2. \u5220\u9664\u201c\u77ff\u7269\u7c7b\u578b\u201d\u5217\u4e2d\u503c\u4e3aE\u7684\u884c
\ndf = df[df["\u77ff\u7269\u7c7b\u578b"] != "E"]<\/p>\n# 3. \u5220\u9664\u201c\u5e8f\u53f7\u201d\u5217
\ndf = df.drop(columns=["\u5e8f\u53f7"])<\/p>\n#4\u3001\u628a\u77ff\u7269\u7c7b\u578b\u503c\u6620\u5c04\u4e3a\u6570\u5b57
\nmapping = {'A': 0, 'B': 1, 'C': 2, 'D': 3} # \u5b9a\u4e49\u6620\u5c04\u89c4\u5219
\ndf["\u77ff\u7269\u7c7b\u578b"] = df["\u77ff\u7269\u7c7b\u578b"].map(mapping) # \u5e94\u7528\u6620\u5c04<\/p>\n# 5. \u5c06\u6240\u6709\u5355\u5143\u683c\u503c\u8f6c\u4e3a\u6570\u503c\u7c7b\u578b，\u5f02\u5e38\u6570\u636e\u8f6c\u4e3anan
\n# \u904d\u5386\u6240\u6709\u5217\u8fdb\u884c\u7c7b\u578b\u8f6c\u6362
\nfor col in df.columns:
\n # pd.to_numeric：\u81ea\u52a8\u5c06\u975e\u6570\u503c\u578b\u6570\u636e\u8f6c\u4e3anan
\n # errors='coerce'：\u5f3a\u5236\u5c06\u65e0\u6cd5\u8f6c\u6362\u7684\u503c\u8f6c\u4e3anan
\n df[col] = pd.to_numeric(df[col], errors='coerce')<\/p>\n#\u6807\u51c6\u5316\u6570\u636e
\nx = df.drop(columns=["\u77ff\u7269\u7c7b\u578b"]) # \u7279\u5f81
\ny = df["\u77ff\u7269\u7c7b\u578b"] # \u6807\u7b7e（\u5206\u7c7b\u76ee\u6807）
\nscaler = StandardScaler()
\nx_z=scaler.fit_transform(x)
\nx=pd.DataFrame(x_z,columns=x.columns)#\u6807\u51c6\u5316\u5f97\u5230\u7684\u6570\u7ec4\u8f6c\u56de\u8868\u683c<\/p>\n#7\u3001\u5212\u5206\u8bad\u7ec3\u96c6\u548c\u6d4b\u8bd5\u96c6
\n'''\u5148\u5212\u5206\u518d\u5904\u7406\u7a7a\u7f3a\u503c，\u907f\u514d\u6570\u636e\u6cc4\u9732'''
\n# \u5212\u5206：\u6d4b\u8bd5\u96c6\u5360\u6bd420%，\u968f\u673a\u79cd\u5b50\u56fa\u5b9a\u4fdd\u8bc1\u7ed3\u679c\u53ef\u590d\u73b0
\nx_train, x_test, y_train, y_test = train_test_split(
\n x, y, test_size=0.2, random_state=4, stratify=y # stratify=y：\u4fdd\u8bc1\u8bad\u7ec3\/\u6d4b\u8bd5\u96c6\u6807\u7b7e\u5206\u5e03\u4e00\u81f4
\n)<\/p>\n#8\u3001\u7a7a\u7f3a\u503c\u5904\u7406<\/p>\n
#\u5220\u9664\u7a7a\u7f3a\u503c\u7684\u884c
\n# x_train_fill,y_train_fill=fill.drop_null_train(x_train,y_train)
\n# x_test_fill,y_test_fill=fill.drop_null_test(x_test,y_test)<\/p>\n# #\u586b\u5145\u4f17\u6570
\n# x_train_fill,y_train_fill=fill.mode_train_fill(x_train,y_train)
\n# x_test_fill,y_test_fill=fill.mode_test_fill(x_train_fill,y_train_fill,x_test,y_test)<\/p>\n# # \u586b\u5145\u5747\u503c\u8c03\u7528（\u548c\u4f60\u4f17\u6570\u8c03\u7528\u683c\u5f0f\u5b8c\u5168\u4e00\u81f4）
\n# x_train_fill,y_train_fill=fill.mean_train_fill(x_train,y_train)
\n# x_test_fill,y_test_fill=fill.mean_test_fill(x_train_fill,y_train_fill,x_test,y_test)<\/p>\n# # \u586b\u5145\u4e2d\u4f4d\u6570\u8c03\u7528（\u548c\u4f60\u4f17\u6570\/\u5747\u503c\u8c03\u7528\u683c\u5f0f\u5b8c\u5168\u4e00\u81f4）
\n# x_train_fill,y_train_fill=fill.median_train_fill(x_train,y_train)
\n# x_test_fill,y_test_fill=fill.median_test_fill(x_train_fill,y_train_fill,x_test,y_test)<\/p>\n# #\u7ebf\u6027\u56de\u5f52\u586b\u5145
\n# x_train_fill,y_train_fill=fill.lr_train_fill(x_train,y_train)
\n# x_test_fill,y_test_fill=fill.lr_test_fill(x_train_fill,y_train_fill,x_test,y_test)<\/p>\n#\u968f\u673a\u68ee\u6797<\/p>\n
x_train_fill,y_train_fill=fill.rf_train_fill(x_train,y_train)
\nx_test_fill,y_test_fill=fill.rf_test_fill(x_train_fill,y_train_fill,x_test,y_test)<\/p>\n#9\u3001\u5747\u8861\u5904\u7406
\noversampler=SMOTE(k_neighbors=1,random_state=50)
\nos_x_train,os_y_train=oversampler.fit_resample(x_train_fill,y_train_fill)<\/p>\n#10\u3001\u4fdd\u5b58\u6570\u636e\u6e05\u6d17\u7684\u7ed3\u679c
\n'''\u6570\u636e\u4fdd\u5b58\u4e3aexcel\u6587\u4ef6'''
\ndata_train = pd.concat([os_y_train,os_x_train],axis=1).sample(frac=1, random_state=4)
\n#sample() \u65b9\u6cd5\u7528\u4e8e\u4eceDataFrame\u4e2d\u968f\u673a\u62bd\u53d6\u884c\u3002frac:\u8868\u793a\u62bd\u53d6\u884c\u7684\u6bd4\u4f8b\u3002
\ndata_test = pd.concat([y_test_fill,x_test_fill],axis=1)
\n#\u6d4b\u8bd5\u96c6\u4e0d\u7528\u4f20\u5165\u6a21\u578b\u8bad\u7ec3， \u65e0\u9700\u6253\u4e71\u987a\u5e8f\u3002<\/p>\ndata_train.to_excel(r'.\/\/temp_data\/\/\u8bad\u7ec3\u6570\u636e\u96c6[\u968f\u673a\u68ee\u6797].xlsx',index=False)
\ndata_test.to_excel(r'.\/\/temp_data\/\/\u6d4b\u8bd5\u6570\u636e\u96c6[\u968f\u673a\u68ee\u6797].xlsx',index=False)<\/p>\n\u5bfc\u5165\u586b\u5145\u51fd\u6570\u4ee3\u7801：<\/p>\n
import pandas as pd
\n#\u5220\u9664\u7a7a\u7f3a\u503c\u884c
\n'''\u8fd9\u91cc\u5206\u4e3a\u4e24\u4e2a\u51fd\u6570，\u5185\u5bb9\u76f8\u540c\u662f\u4e3a\u4e86：
\n\u6d4b\u8bd5\u96c6\u5e94\u8be5\u57fa\u4e8e\u8bad\u7ec3\u96c6\u6765\u5904\u7406\u7f3a\u5931\u503c，\u907f\u514d\u6570\u636e\u6cc4\u9732
\n\u53ea\u4e0d\u8fc7\u5220\u9664\u884c\u7684\u65b9\u6cd5\u7528\u4e0d\u5230'''
\ndef drop_null_train(x,y):
\n data=pd.concat([x,y],axis=1)
\n data=data.reset_index(drop=True)#\u91cd\u8981，\u6570\u636e\u4e4b\u524d\u6392\u5e8f\u7b5b\u9009\u7b49\u5404\u79cd\u64cd\u4f5c\u4f1a\u6253\u4e71\u7d22\u5f15，\u4e0d\u91cd\u7f6e\u540e\u7eed\u62fc\u63a5\u4f1a\u9519\u4e71
\n df_filled=data.dropna()
\n return df_filled.drop("\u77ff\u7269\u7c7b\u578b",axis=1),df_filled.\u77ff\u7269\u7c7b\u578b
\ndef drop_null_test(x,y):
\n data=pd.concat([x,y],axis=1)
\n data=data.reset_index(drop=True)
\n df_filled=data.dropna()
\n return df_filled.drop("\u77ff\u7269\u7c7b\u578b",axis=1),df_filled.\u77ff\u7269\u7c7b\u578b<\/p>\n#\u586b\u5145\u4f17\u6570
\ndef mode_method(data):
\n fill_values=data.apply(lambda x:x.mode().iloc[0] if len(x.mode())>0 else None)
\n a=data.mode()
\n return data.fillna(fill_values)<\/p>\ndef mode_train_fill(x_train,y_train):
\n data=pd.concat([x_train,y_train],axis=1)
\n data=data.reset_index(drop=True)
\n A=data[data["\u77ff\u7269\u7c7b\u578b"]==0]
\n B=data[data["\u77ff\u7269\u7c7b\u578b"]==1]
\n C=data[data["\u77ff\u7269\u7c7b\u578b"]==2]
\n D=data[data["\u77ff\u7269\u7c7b\u578b"]==3]<\/p>\nA=mode_method(A)
\n B=mode_method(B)
\n C=mode_method(C)
\n D=mode_method(D)
\n df_filled=pd.concat([A,B,C,D])
\n df_filled=df_filled.reset_index(drop=True)
\n return df_filled.drop("\u77ff\u7269\u7c7b\u578b",axis=1),df_filled.\u77ff\u7269\u7c7b\u578b<\/p>\ndef mode_method_test(data,data1):
\n fill_values = data.apply(lambda x: x.mode().iloc[0] if len(x.mode()) > 0 else None)
\n return data1.fillna(fill_values)<\/p>\ndef mode_test_fill(x_train_fill,y_train_fill,x_test,y_test):
\n data = pd.concat([x_train_fill,y_train_fill], axis=1)
\n data = data.reset_index(drop=True)
\n data1 = pd.concat([x_test,y_test], axis=1)
\n data1 = data1.reset_index(drop=True)
\n A = data[data["\u77ff\u7269\u7c7b\u578b"] == 0]
\n B = data[data["\u77ff\u7269\u7c7b\u578b"] == 1]
\n C = data[data["\u77ff\u7269\u7c7b\u578b"] == 2]
\n D = data[data["\u77ff\u7269\u7c7b\u578b"] == 3]
\n A1 = data1[data1["\u77ff\u7269\u7c7b\u578b"] == 0]
\n B1 = data1[data1["\u77ff\u7269\u7c7b\u578b"] == 1]
\n C1 = data1[data1["\u77ff\u7269\u7c7b\u578b"] == 2]
\n D1 = data1[data1["\u77ff\u7269\u7c7b\u578b"] == 3]
\n A1 = mode_method_test(A,A1)
\n B1 = mode_method_test(B,B1)
\n C1 = mode_method_test(C,C1)
\n D1 = mode_method_test(D,D1)
\n df_filled = pd.concat([A1, B1, C1, D1])
\n df_filled = df_filled.reset_index(drop=True)
\n return df_filled.drop("\u77ff\u7269\u7c7b\u578b", axis=1), df_filled.\u77ff\u7269\u7c7b\u578b<\/p>\n#\u586b\u5145\u5747\u503c
\n# \u586b\u5145\u5747\u503c
\ndef mean_method(data):
\n fill_values=data.apply(lambda x:x.mean() if len(x.dropna())>0 else None) # \u8ba1\u7b97\u5217\u5747\u503c（\u5254\u9664\u7a7a\u503c）
\n return data.fillna(fill_values)<\/p>\ndef mean_train_fill(x_train,y_train):
\n data=pd.concat([x_train,y_train],axis=1)
\n data=data.reset_index(drop=True)
\n A=data[data["\u77ff\u7269\u7c7b\u578b"]==0]
\n B=data[data["\u77ff\u7269\u7c7b\u578b"]==1]
\n C=data[data["\u77ff\u7269\u7c7b\u578b"]==2]
\n D=data[data["\u77ff\u7269\u7c7b\u578b"]==3]<\/p>\nA=mean_method(A)
\n B=mean_method(B)
\n C=mean_method(C)
\n D=mean_method(D)
\n df_filled=pd.concat([A,B,C,D])
\n df_filled=df_filled.reset_index(drop=True)
\n return df_filled.drop("\u77ff\u7269\u7c7b\u578b",axis=1),df_filled.\u77ff\u7269\u7c7b\u578b<\/p>\ndef mean_method_test(data,data1):
\n fill_values = data.apply(lambda x:x.mean() if len(x.dropna())>0 else None) # \u7528\u8bad\u7ec3\u96c6\u5206\u7ec4\u5747\u503c
\n return data1.fillna(fill_values)<\/p>\ndef mean_test_fill(x_train_fill,y_train_fill,x_test,y_test):
\n data = pd.concat([x_train_fill,y_train_fill], axis=1)
\n data = data.reset_index(drop=True)
\n data1 = pd.concat([x_test,y_test], axis=1)
\n data1 = data1.reset_index(drop=True)
\n A = data[data["\u77ff\u7269\u7c7b\u578b"] == 0]
\n B = data[data["\u77ff\u7269\u7c7b\u578b"] == 1]
\n C = data[data["\u77ff\u7269\u7c7b\u578b"] == 2]
\n D = data[data["\u77ff\u7269\u7c7b\u578b"] == 3]
\n A1 = data1[data1["\u77ff\u7269\u7c7b\u578b"] == 0]
\n B1 = data1[data1["\u77ff\u7269\u7c7b\u578b"] == 1]
\n C1 = data1[data1["\u77ff\u7269\u7c7b\u578b"] == 2]
\n D1 = data1[data1["\u77ff\u7269\u7c7b\u578b"] == 3]
\n A1 = mean_method_test(A,A1)
\n B1 = mean_method_test(B,B1)
\n C1 = mean_method_test(C,C1)
\n D1 = mean_method_test(D,D1)
\n df_filled = pd.concat([A1, B1, C1, D1])
\n df_filled = df_filled.reset_index(drop=True)
\n return df_filled.drop("\u77ff\u7269\u7c7b\u578b", axis=1), df_filled.\u77ff\u7269\u7c7b\u578b<\/p>\n#\u586b\u5145\u4e2d\u4f4d\u6570
\n# \u586b\u5145\u4e2d\u4f4d\u6570
\ndef median_method(data):
\n fill_values=data.apply(lambda x:x.median() if len(x.dropna())>0 else None) # \u8ba1\u7b97\u5217\u4e2d\u4f4d\u6570（\u5254\u9664\u7a7a\u503c）
\n return data.fillna(fill_values)<\/p>\ndef median_train_fill(x_train,y_train):
\n data=pd.concat([x_train,y_train],axis=1)
\n data=data.reset_index(drop=True)
\n A=data[data["\u77ff\u7269\u7c7b\u578b"]==0]
\n B=data[data["\u77ff\u7269\u7c7b\u578b"]==1]
\n C=data[data["\u77ff\u7269\u7c7b\u578b"]==2]
\n D=data[data["\u77ff\u7269\u7c7b\u578b"]==3]<\/p>\nA=median_method(A)
\n B=median_method(B)
\n C=median_method(C)
\n D=median_method(D)
\n df_filled=pd.concat([A,B,C,D])
\n df_filled=df_filled.reset_index(drop=True)
\n return df_filled.drop("\u77ff\u7269\u7c7b\u578b",axis=1),df_filled.\u77ff\u7269\u7c7b\u578b<\/p>\ndef median_method_test(data,data1):
\n fill_values = data.apply(lambda x:x.median() if len(x.dropna())>0 else None) # \u7528\u8bad\u7ec3\u96c6\u5206\u7ec4\u4e2d\u4f4d\u6570
\n return data1.fillna(fill_values)<\/p>\ndef median_test_fill(x_train_fill,y_train_fill,x_test,y_test):
\n data = pd.concat([x_train_fill,y_train_fill], axis=1)
\n data = data.reset_index(drop=True)
\n data1 = pd.concat([x_test,y_test], axis=1)
\n data1 = data1.reset_index(drop=True)
\n A = data[data["\u77ff\u7269\u7c7b\u578b"] == 0]
\n B = data[data["\u77ff\u7269\u7c7b\u578b"] == 1]
\n C = data[data["\u77ff\u7269\u7c7b\u578b"] == 2]
\n D = data[data["\u77ff\u7269\u7c7b\u578b"] == 3]
\n A1 = data1[data1["\u77ff\u7269\u7c7b\u578b"] == 0]
\n B1 = data1[data1["\u77ff\u7269\u7c7b\u578b"] == 1]
\n C1 = data1[data1["\u77ff\u7269\u7c7b\u578b"] == 2]
\n D1 = data1[data1["\u77ff\u7269\u7c7b\u578b"] == 3]
\n A1 = median_method_test(A,A1)
\n B1 = median_method_test(B,B1)
\n C1 = median_method_test(C,C1)
\n D1 = median_method_test(D,D1)
\n df_filled = pd.concat([A1, B1, C1, D1])
\n df_filled = df_filled.reset_index(drop=True)
\n return df_filled.drop("\u77ff\u7269\u7c7b\u578b", axis=1), df_filled.\u77ff\u7269\u7c7b\u578b<\/p>\n#\u7ebf\u6027\u56de\u5f52\u9884\u6d4b\u7f3a\u5931\u503c\u8fdb\u884c\u586b\u5145
\nfrom sklearn.linear_model import LinearRegression<\/p>\n# def lr_train_fill(x_train,y_train):
\n# train_data_all=pd.concat([x_train,y_train],axis=1)
\n# train_data_all=train_data_all.reset_index(drop=True)
\n# x_train=train_data_all.drop("\u77ff\u7269\u7c7b\u578b",axis=1)
\n# null_num=x_train.isnull().sum()
\n# null_num_sorted=null_num.sort_values(ascending=True)
\n#
\n# filling_feature=[]
\n# for i in null_num_sorted.index:
\n# filling_feature.append(i)
\n# if null_num_sorted[i]!=0:
\n# x=x_train[filling_feature].drop(i,axis=1)
\n# y=x_train[i]
\n# row_numbers_mg_null=x_train[x_train[i].isnull()].index.tolist()
\n# train=x.drop(row_numbers_mg_null)
\n# test=y.drop(row_numbers_mg_null)
\n# x_test=x.iloc[row_numbers_mg_null]
\n# regr=LinearRegression()
\n# regr.fit(train,test)
\n# y_pred=regr.predict(x_test)
\n# x_train.iloc[row_numbers_mg_null]=y_pred
\n# print("\u8bad\u7ec3\u96c6{}\u5217\u586b\u5145\u5b8c\u6bd5".format(i))
\n# return x_train,train_data_all.\u77ff\u7269\u7c7b\u578b<\/p>\n# def lr_train_fill(x_train,y_train):
\n# train_data_all = pd.concat([x_train, y_train], axis=1)
\n# train_data_all = train_data_all.reset_index(drop=True)
\n# train_data_X = train_data_all.drop('\u77ff\u7269\u7c7b\u578b', axis=1)
\n# null_num = train_data_X.isnull().sum() # \u67e5\u770b\u6bcf\u4e2a\u79cd\u4e2d\u5b58\u5728\u7a7a\u6570\u636e\u7684\u4e2a\u6570
\n# null_num_sorted = null_num.sort_values(ascending=True) # \u5c06\u7a7a\u6570\u636e\u7684\u7c7b\u522b\u4ece\u5c0f\u5230
\n#
\n# filling_feature = [] # \u7528\u6765\u5b58\u50a8\u9700\u8981\u4f20\u5165\u6a21\u578b\u7684\u7279\u5f81\u540d\u79f0
\n#
\n# for i in null_num_sorted.index:
\n# filling_feature.append(i)
\n# if null_num_sorted[i] != 0: # \u5f53\u524d\u7279\u5f81\u662f\u5426\u6709\u7a7a\u7f3a\u7684\u5185\u5bb9\u3002\u7528\u6765\u5224\u65ad\u662f\u5426\u5f00\u59cb\u8bad\u7ec3\u6a21\u578b
\n#
\n# X = train_data_X[filling_feature].drop(i, axis=1) # \u6784\u5efa\u8bad\u7ec3\u96c6
\n# y = train_data_X[i] # \u6784\u5efa\u6d4b\u8bd5\u96c6
\n#
\n# row_numbers_mg_null = train_data_X[train_data_X[i].isnull()].index.tolist()
\n# X_train = X.drop(row_numbers_mg_null) # \u975e\u7a7a\u7684\u6570\u636e\u4f5c\u4e3a\u8bad\u7ec3\u6570\u636e\u96c6
\n# y_train = y.drop(row_numbers_mg_null) # \u975e\u7a7a\u7684\u6807\u7b7e\u4f5c\u4e3a\u8bad\u7ec3\u6807\u7b7e
\n# X_test = X.iloc[row_numbers_mg_null] # \u7a7a\u7684\u6570\u636e\u4f5c\u4e3a\u6d4b\u8bd5\u6570\u636e\u96c6
\n#
\n# regr = LinearRegression() # \u521b\u5efa\u7ebf\u6027\u56de\u5f52\u6a21\u578b
\n# regr.fit(X_train, y_train) # \u8bad\u7ec3\u6a21\u578b
\n# y_pred = regr.predict(X_test) # \u4f7f\u7528\u6a21\u578b\u8fdb\u884c\u9884\u6d4b
\n# train_data_X.loc[row_numbers_mg_null] = y_pred # pandas.loc[3,4
\n# print('\u5b8c\u6210\u8bad\u7ec3\u6570\u636e\u4e2d\u7684{}\u5217\u6570\u636e\u7684\u586b\u5145'.format(i))
\n# return train_data_X,train_data_X.\u77ff\u7269\u7c7b\u578b<\/p>\ndef lr_train_fill(x_train, y_train):
\n # \u6838\u5fc3\u4fee\u590d：\u62fc\u63a5\u524d\u91cd\u7f6e\u7d22\u5f15，\u786e\u4fddx_train\u548cy_train\u7d22\u5f15\u5b8c\u5168\u5bf9\u9f50
\n x_train = x_train.reset_index(drop=True) # \u91cd\u7f6e\u7279\u5f81\u7d22\u5f15，\u4e22\u5f03\u539f\u7d22\u5f15
\n y_train = y_train.reset_index(drop=True) # \u91cd\u7f6e\u6807\u7b7e\u7d22\u5f15，\u4e22\u5f03\u539f\u7d22\u5f15<\/p>\n# \u518d\u62fc\u63a5，\u6b64\u65f6\u7d22\u5f15\u5b8c\u5168\u4e00\u81f4，\u4e0d\u4f1a\u51fa\u73b0\u6574\u884c\u7a7a
\n train_data_all = pd.concat([x_train, y_train], axis=1)
\n # \u540e\u7eed\u4ee3\u7801\u4e0d\u53d8（\u53ef\u4fdd\u7559\u539freset_index，\u4e0d\u5f71\u54cd）
\n train_data_all = train_data_all.reset_index(drop=True)
\n train_data_X = train_data_all.drop('\u77ff\u7269\u7c7b\u578b', axis=1)
\n null_num = train_data_X.isnull().sum()
\n null_num_sorted = null_num.sort_values(ascending=True)<\/p>\nfilling_feature = []<\/p>\n
for i in null_num_sorted.index:
\n filling_feature.append(i)
\n if null_num_sorted[i] != 0:<\/p>\nX = train_data_X[filling_feature].drop(i, axis=1)
\n y = train_data_X[i]<\/p>\nrow_numbers_mg_null = train_data_X[train_data_X[i].isnull()].index.tolist()
\n # \u4fee\u590d\u53d8\u91cf\u540d\u51b2\u7a81：\u5c06\u5c40\u90e8X_train\/y_train\u6539\u4e3aX_train_col\/y_train_col
\n X_train_col = X.drop(row_numbers_mg_null)
\n y_train_col = y.drop(row_numbers_mg_null)
\n X_test = X.iloc[row_numbers_mg_null]<\/p>\n# \u65b0\u589e：\u6821\u9a8c\u8bad\u7ec3\u96c6\u662f\u5426\u4e3a\u7a7a，\u907f\u514d\u62a5\u9519
\n if not X_train_col.empty and not y_train_col.empty:
\n regr = LinearRegression()
\n regr.fit(X_train_col, y_train_col)
\n y_pred = regr.predict(X_test)
\n # \u4fee\u590d：\u4ec5\u586b\u5145\u5f53\u524d\u5217i\u7684\u7a7a\u503c，\u800c\u975e\u6574\u884c
\n train_data_X.loc[row_numbers_mg_null, i] = y_pred
\n print('\u5b8c\u6210\u8bad\u7ec3\u6570\u636e\u4e2d\u7684{}\u5217\u6570\u636e\u7684\u586b\u5145'.format(i))
\n else:
\n print(f"\u8b66\u544a：\u5217{i}\u5220\u9664\u7a7a\u503c\u540e\u65e0\u6709\u6548\u6837\u672c，\u8df3\u8fc7\u586b\u5145")<\/p>\n# \u4fee\u590d\u8fd4\u56de\u503c：train_data_X\u4e0d\u542b\u77ff\u7269\u7c7b\u578b，\u9700\u4ecetrain_data_all\u63d0\u53d6
\n return train_data_X, train_data_all["\u77ff\u7269\u7c7b\u578b"]<\/p>\n# def lr_test_fill(x_train_fill,y_train_fill,x_test,y_test):
\n# train_data_all=pd.concat([x_train_fill,y_train_fill],axis=1)
\n# train_data_all=train_data_all.reset_index(drop=True)
\n#
\n# test_data_all=pd.concat([x_test,y_test],axis=1)
\n# test_data_all=test_data_all.reset_index(drop=True)
\n#
\n# train_data_x=train_data_all.drop("\u77ff\u7269\u7c7b\u578b",axis=1)
\n# test_data_x=test_data_all.drop("\u77ff\u7269\u7c7b\u578b",axis=1)
\n# null_num=test_data_x.isnull().sum()
\n# null_sum_sorted=null_num.sort_values(ascending=True)
\n#
\n# filling_feature=[]
\n# for i in null_sum_sorted.index:
\n# filling_feature.append(i)
\n# if null_sum_sorted[i]!=0:
\n# X_train = train_data_x[filling_feature].drop(i, axis=1)
\n# y_train = train_data_x[i]
\n# X_test = test_data_x[filling_feature].drop(i, axis=1)
\n# row_numbers_mg_null = test_data_x[test_data_x[i].isnull()].index.tolist() #
\n# X_test = X_test.iloc[row_numbers_mg_null] # \u7a7a\u7684\u6570\u636e\u4f5c\u4e3a\u6d4b\u8bd5\u6570\u636e\u96c6
\n#
\n# regr = LinearRegression() # \u521b\u5efa\u968f\u673a\u68ee\u6797\u56de\u5f52\u6a21\u578b
\n# regr.fit(X_train, y_train) # \u4f7f\u7528\u6a21\u578b\u8fdb\u884c\u8bad\u7ec3
\n# y_pred = regr.predict(X_test) # \u4f7f\u7528\u6a21\u578b\u8fdb\u884c\u9884\u6d4b
\n# test_data_x.iloc[row_numbers_mg_null]=y_pred
\n# return test_data_x,test_data_all.\u77ff\u7269\u7c7b\u578b<\/p>\nfrom sklearn.linear_model import LinearRegression # \u786e\u4fdd\u5bfc\u5165\u7ebf\u6027\u56de\u5f52<\/p>\n
def lr_test_fill(x_train_fill, y_train_fill, x_test, y_test):
\n # \u6838\u5fc3\u4fee\u590d1：\u62fc\u63a5\u524d\u5f3a\u5236\u91cd\u7f6e\u7d22\u5f15，\u786e\u4fdd\u7279\u5f81\u548c\u6807\u7b7e\u7d22\u5f15\u5b8c\u5168\u5bf9\u9f50
\n x_train_fill = x_train_fill.reset_index(drop=True)
\n y_train_fill = y_train_fill.reset_index(drop=True)
\n x_test = x_test.reset_index(drop=True)
\n y_test = y_test.reset_index(drop=True)<\/p>\n# \u62fc\u63a5\u8bad\u7ec3\u96c6（\u7528\u4e8e\u8bad\u7ec3\u6a21\u578b）\u548c\u6d4b\u8bd5\u96c6（\u5f85\u586b\u5145）
\n train_data_all = pd.concat([x_train_fill, y_train_fill], axis=1)
\n train_data_all = train_data_all.reset_index(drop=True)
\n test_data_all = pd.concat([x_test, y_test], axis=1)
\n test_data_all = test_data_all.reset_index(drop=True)<\/p>\n# \u5206\u79bb\u7279\u5f81（\u5220\u9664\u77ff\u7269\u7c7b\u578b\u5217）
\n train_data_x = train_data_all.drop("\u77ff\u7269\u7c7b\u578b", axis=1)
\n test_data_x = test_data_all.drop("\u77ff\u7269\u7c7b\u578b", axis=1)<\/p>\n# \u6309\u7a7a\u503c\u6570\u91cf\u6392\u5e8f，\u9010\u5217\u586b\u5145
\n null_num = test_data_x.isnull().sum()
\n null_sum_sorted = null_num.sort_values(ascending=True)<\/p>\nfilling_feature = []
\n for i in null_sum_sorted.index:
\n filling_feature.append(i)
\n if null_sum_sorted[i] != 0: # \u4ec5\u5904\u7406\u6709\u7a7a\u503c\u7684\u5217
\n # \u6784\u5efa\u8bad\u7ec3\u96c6（\u7528\u5df2\u586b\u5145\u597d\u7684\u8bad\u7ec3\u96c6\u6570\u636e）
\n X_train = train_data_x[filling_feature].drop(i, axis=1)
\n y_train = train_data_x[i]<\/p>\n# \u6784\u5efa\u6d4b\u8bd5\u96c6（\u5f85\u586b\u5145\u7684\u6d4b\u8bd5\u96c6\u7279\u5f81）
\n X_test = test_data_x[filling_feature].drop(i, axis=1)
\n row_numbers_mg_null = test_data_x[test_data_x[i].isnull()].index.tolist()
\n X_test_null = X_test.iloc[row_numbers_mg_null] # \u4ec5\u63d0\u53d6\u7a7a\u503c\u884c\u7684\u7279\u5f81
\n regr = LinearRegression()
\n regr.fit(X_train, y_train)
\n y_pred = regr.predict(X_test_null)<\/p>\n# \u6838\u5fc3\u4fee\u590d3：\u4ec5\u586b\u5145\u5f53\u524d\u5217i\u7684\u7a7a\u503c，\u800c\u975e\u6574\u884c
\n test_data_x.loc[row_numbers_mg_null, i] = y_pred<\/p>\n# \u8fd4\u56de\u586b\u5145\u540e\u7684\u6d4b\u8bd5\u96c6\u7279\u5f81+\u539f\u59cb\u6807\u7b7e（\u77ff\u7269\u7c7b\u578b）
\n return test_data_x, test_data_all["\u77ff\u7269\u7c7b\u578b"]<\/p>\n#\u968f\u673a\u68ee\u6797
\nimport pandas as pd
\nfrom sklearn.ensemble import RandomForestRegressor # \u5bfc\u5165\u968f\u673a\u68ee\u6797\u56de\u5f52\u5668<\/p>\n# ———————- \u968f\u673a\u68ee\u6797\u586b\u5145\u8bad\u7ec3\u96c6 ———————-
\ndef rf_train_fill(x_train, y_train):
\n # 1. \u7d22\u5f15\u5bf9\u9f50：\u62fc\u63a5\u524d\u91cd\u7f6e\u6240\u6709\u8f93\u5165\u7d22\u5f15
\n x_train = x_train.reset_index(drop=True)
\n y_train = y_train.reset_index(drop=True)<\/p>\n# 2. \u62fc\u63a5\u7279\u5f81+\u6807\u7b7e，\u5206\u79bb\u77ff\u7269\u7c7b\u578b
\n train_data_all = pd.concat([x_train, y_train], axis=1)
\n train_data_all = train_data_all.reset_index(drop=True)
\n train_data_x = train_data_all.drop('\u77ff\u7269\u7c7b\u578b', axis=1)<\/p>\n# 3. \u6309\u7a7a\u503c\u6570\u91cf\u4ece\u5c0f\u5230\u5927\u6392\u5e8f，\u9010\u5217\u586b\u5145
\n null_num = train_data_x.isnull().sum()
\n null_num_sorted = null_num.sort_values(ascending=True)
\n filling_feature = []<\/p>\nfor i in null_num_sorted.index:
\n filling_feature.append(i)
\n if null_num_sorted[i] != 0: # \u4ec5\u5904\u7406\u6709\u7a7a\u503c\u7684\u5217
\n # \u6784\u5efa\u5f53\u524d\u5217\u7684\u8bad\u7ec3\u7279\u5f81\u548c\u6807\u7b7e
\n X = train_data_x[filling_feature].drop(i, axis=1)
\n y = train_data_x[i]
\n row_numbers_mg_null = train_data_x[train_data_x[i].isnull()].index.tolist()<\/p>\n# \u5206\u79bb\u975e\u7a7a\u8bad\u7ec3\u96c6\u548c\u5f85\u586b\u5145\u6d4b\u8bd5\u96c6
\n X_train_col = X.drop(row_numbers_mg_null)
\n y_train_col = y.drop(row_numbers_mg_null)
\n X_test_col = X.iloc[row_numbers_mg_null]<\/p>\n# \u975e\u7a7a\u6821\u9a8c：\u907f\u514d\u6a21\u578b\u8bad\u7ec3\u62a5\u9519
\n if not X_train_col.empty and not y_train_col.empty:
\n # \u6838\u5fc3\u66ff\u6362：\u7ebf\u6027\u56de\u5f52 \u2192 \u968f\u673a\u68ee\u6797\u56de\u5f52
\n rf = RandomForestRegressor(
\n n_estimators=100, # \u51b3\u7b56\u6811\u6570\u91cf（\u53ef\u8c03）
\n random_state=42, # \u56fa\u5b9a\u968f\u673a\u79cd\u5b50，\u4fdd\u8bc1\u7ed3\u679c\u53ef\u590d\u73b0
\n n_jobs=-1 # \u591a\u7ebf\u7a0b\u52a0\u901f
\n )
\n rf.fit(X_train_col, y_train_col) # \u8bad\u7ec3\u968f\u673a\u68ee\u6797
\n y_pred = rf.predict(X_test_col) # \u9884\u6d4b\u7a7a\u503c<\/p>\n# \u7cbe\u51c6\u586b\u5145：\u4ec5\u66f4\u65b0\u5f53\u524d\u5217i\u7684\u7a7a\u503c
\n train_data_x.loc[row_numbers_mg_null, i] = y_pred
\n print(f'\u5b8c\u6210\u8bad\u7ec3\u6570\u636e\u4e2d\u3010{i}\u3011\u5217\u7684\u968f\u673a\u68ee\u6797\u586b\u5145')
\n else:
\n print(f'\u8b66\u544a：\u8bad\u7ec3\u96c6\u3010{i}\u3011\u5217\u5220\u9664\u7a7a\u503c\u540e\u65e0\u6709\u6548\u6837\u672c，\u8df3\u8fc7\u586b\u5145')<\/p>\n# \u8fd4\u56de\u586b\u5145\u540e\u7684\u7279\u5f81+\u539f\u59cb\u6807\u7b7e
\n return train_data_x, train_data_all["\u77ff\u7269\u7c7b\u578b"]<\/p>\n# ———————- \u968f\u673a\u68ee\u6797\u586b\u5145\u6d4b\u8bd5\u96c6 ———————-
\ndef rf_test_fill(x_train_fill, y_train_fill, x_test, y_test):
\n # 1. \u7d22\u5f15\u5bf9\u9f50：\u62fc\u63a5\u524d\u91cd\u7f6e\u6240\u6709\u8f93\u5165\u7d22\u5f15
\n x_train_fill = x_train_fill.reset_index(drop=True)
\n y_train_fill = y_train_fill.reset_index(drop=True)
\n x_test = x_test.reset_index(drop=True)
\n y_test = y_test.reset_index(drop=True)<\/p>\n# 2. \u62fc\u63a5\u8bad\u7ec3\u96c6\/\u6d4b\u8bd5\u96c6，\u5206\u79bb\u77ff\u7269\u7c7b\u578b
\n train_data_all = pd.concat([x_train_fill, y_train_fill], axis=1)
\n train_data_all = train_data_all.reset_index(drop=True)
\n test_data_all = pd.concat([x_test, y_test], axis=1)
\n test_data_all = test_data_all.reset_index(drop=True)<\/p>\ntrain_data_x = train_data_all.drop("\u77ff\u7269\u7c7b\u578b", axis=1)
\n test_data_x = test_data_all.drop("\u77ff\u7269\u7c7b\u578b", axis=1)<\/p>\n# 3. \u6309\u7a7a\u503c\u6570\u91cf\u6392\u5e8f，\u9010\u5217\u586b\u5145
\n null_num = test_data_x.isnull().sum()
\n null_sum_sorted = null_num.sort_values(ascending=True)
\n filling_feature = []<\/p>\nfor i in null_sum_sorted.index:
\n filling_feature.append(i)
\n if null_sum_sorted[i] != 0: # \u4ec5\u5904\u7406\u6709\u7a7a\u503c\u7684\u5217
\n # \u7528\u8bad\u7ec3\u96c6\u6784\u5efa\u6a21\u578b\u7279\u5f81，\u6d4b\u8bd5\u96c6\u6784\u5efa\u5f85\u586b\u5145\u7279\u5f81
\n X_train = train_data_x[filling_feature].drop(i, axis=1)
\n y_train = train_data_x[i]
\n X_test = test_data_x[filling_feature].drop(i, axis=1)<\/p>\nrow_numbers_mg_null = test_data_x[test_data_x[i].isnull()].index.tolist()
\n X_test_null = X_test.iloc[row_numbers_mg_null]<\/p>\n# \u975e\u7a7a\u6821\u9a8c
\n if not X_train.empty and not y_train.empty and not X_test_null.empty:
\n # \u6838\u5fc3\u66ff\u6362：\u7ebf\u6027\u56de\u5f52 \u2192 \u968f\u673a\u68ee\u6797\u56de\u5f52
\n rf = RandomForestRegressor(
\n n_estimators=100,
\n random_state=42,
\n n_jobs=-1
\n )
\n rf.fit(X_train, y_train) # \u7528\u8bad\u7ec3\u96c6\u8bad\u7ec3
\n y_pred = rf.predict(X_test_null) # \u9884\u6d4b\u6d4b\u8bd5\u96c6\u7a7a\u503c<\/p>\n# \u7cbe\u51c6\u586b\u5145：\u4ec5\u66f4\u65b0\u5f53\u524d\u5217i\u7684\u7a7a\u503c
\n test_data_x.loc[row_numbers_mg_null, i] = y_pred
\n print(f'\u5b8c\u6210\u6d4b\u8bd5\u6570\u636e\u4e2d\u3010{i}\u3011\u5217\u7684\u968f\u673a\u68ee\u6797\u586b\u5145')
\n else:
\n print(f'\u8b66\u544a：\u6d4b\u8bd5\u96c6\u3010{i}\u3011\u5217\u65e0\u6709\u6548\u8bad\u7ec3\/\u6d4b\u8bd5\u6837\u672c，\u8df3\u8fc7\u586b\u5145')<\/p>\n# \u8fd4\u56de\u586b\u5145\u540e\u7684\u6d4b\u8bd5\u96c6\u7279\u5f81+\u539f\u59cb\u6807\u7b7e
\n return test_data_x, test_data_all["\u77ff\u7269\u7c7b\u578b"]<\/p>\n\u9636\u6bb5\u4e8c：\u591a\u6a21\u578b\u5206\u7c7b\u8bad\u7ec3<\/h4>\n
\u57fa\u4e8e\u9884\u5904\u7406\u540e\u76846\u7c7b\u6570\u636e\u96c6，\u5206\u522b\u8bad\u7ec37\u79cd\u5206\u7c7b\u6a21\u578b，\u6db5\u76d6\u4f20\u7edf\u673a\u5668\u5b66\u4e60\u3001\u96c6\u6210\u5b66\u4e60\u3001\u6df1\u5ea6\u5b66\u4e60\u4e09\u5927\u7c7b\u522b，\u786e\u4fdd\u65b9\u6848\u7684\u5168\u9762\u6027\u3002<\/p>\n
2.1 \u4f20\u7edf\u673a\u5668\u5b66\u4e60\u6a21\u578b<\/h5>\n
- \n
- \n
\u903b\u8f91\u56de\u5f52：\u4f5c\u4e3a\u57fa\u7ebf\u6a21\u578b，\u7ed3\u6784\u7b80\u5355\u3001\u8bad\u7ec3\u901f\u5ea6\u5feb，\u9002\u5408\u521d\u6b65\u9a8c\u8bc1\u6570\u636e\u6709\u6548\u6027\u3002\u53c2\u6570\u914d\u7f6e：max_iter=1000（\u786e\u4fdd\u6536\u655b）\u3001random_state=42（\u7ed3\u679c\u53ef\u590d\u73b0），\u7981\u7528\u591a\u7ebf\u7a0b\u907f\u514d\u5185\u5b58\u8fc7\u8f7d\u3002<\/p>\n<\/li>\n
- \n
SVM（\u652f\u6301\u5411\u91cf\u673a）：\u9009\u7528Poly\u6838（\u591a\u9879\u5f0f\u6838），\u80fd\u6355\u6349\u7279\u5f81\u95f4\u7684\u975e\u7ebf\u6027\u5173\u7cfb\u3002\u6838\u5fc3\u95ee\u9898\u89e3\u51b3：\u521d\u59cb\u7528degree=4+probability=True\u65f6，\u56e0\u8ba1\u7b97\u590d\u6742\u5ea6\u6781\u9ad8，\u8bad\u7ec3\u4e2d\u4f4d\u6570\u586b\u5145\u6570\u636e\u96c6\u65f6\u5361\u4f4f，\u4f18\u5316\u540e\u8c03\u6574degree=2\u3001\u5173\u95edprobability，\u540c\u65f6\u589e\u52a0\u7f13\u5b58\u5927\u5c0f（cache_size=2000），\u5927\u5e45\u63d0\u5347\u8bad\u7ec3\u901f\u5ea6\u3002<\/p>\n<\/li>\n<\/ul>\n
2.2 \u96c6\u6210\u5b66\u4e60\u6a21\u578b<\/h5>\n
- \n
- \n
\u968f\u673a\u68ee\u6797：\u96c6\u6210\u591a\u68f5\u51b3\u7b56\u6811，\u6297\u8fc7\u62df\u5408\u80fd\u529b\u5f3a\u3002\u53c2\u6570\u914d\u7f6e：n_estimators=50\u3001max_depth=20\u3001max_features="log2"，\u542f\u7528\u591a\u7ebf\u7a0b（n_jobs=-1）\u52a0\u901f\u8bad\u7ec3\u3002<\/p>\n<\/li>\n
- \n
AdaBoost：\u57fa\u4e8e\u63d0\u5347\u7b56\u7565，\u8fed\u4ee3\u4f18\u5316\u5f31\u5206\u7c7b\u5668\u3002\u4ee5\u6df1\u5ea6\u4e3a2\u7684\u51b3\u7b56\u6811\u4e3a\u57fa\u5206\u7c7b\u5668，n_estimators=200\u3001learning_rate=1.0，\u9002\u914d\u5c0f\u6837\u672c\u6570\u636e\u3002<\/p>\n<\/li>\n
- \n
XGBoost：\u9ad8\u6027\u80fd\u68af\u5ea6\u63d0\u5347\u6811，\u64c5\u957f\u5904\u7406\u590d\u6742\u7279\u5f81\u4ea4\u4e92\u3002\u53c2\u6570\u914d\u7f6e：learning_rate=0.05\u3001max_depth=7\u3001subsample=0.6\u3001colsample_bytree=0.8，\u76ee\u6807\u51fd\u6570\u8bbe\u4e3amulti:softprob（\u9002\u914d\u591a\u5206\u7c7b）\u3002\u5173\u952e\u95ee\u9898\u89e3\u51b3：\u56e0\u6807\u7b7e\u9700\u4e3a\u201c\u4ece0\u5f00\u59cb\u7684\u8fde\u7eed\u6574\u6570\u201d，\u5426\u5219\u62a5\u9519，\u6545\u6dfb\u52a0\u6807\u7b7e\u6620\u5c04\u903b\u8f91，\u5c06\u539f\u59cb\u6807\u7b7e\u52a8\u6001\u8f6c\u4e3a\u8fde\u7eed\u6574\u6570，\u540c\u65f6\u52a8\u6001\u8bbe\u7f6enum_class，\u9002\u914d\u4e0d\u540c\u6570\u636e\u96c6\u7684\u7c7b\u522b\u6570\u3002<\/p>\n<\/li>\n<\/ul>\n
2.3 \u6df1\u5ea6\u5b66\u4e60\u6a21\u578b<\/h5>\n
- \n
- \n
MLP（\u591a\u5c42\u611f\u77e5\u673a）：\u4e09\u5c42\u5168\u8fde\u63a5\u7f51\u7edc，\u9002\u914d\u4e00\u7ef4\u7279\u5f81\u8f93\u5165\u3002\u7ed3\u6784：\u8f93\u5165\u5c42\u2192\u9690\u85cf\u5c42（32\u795e\u7ecf\u5143）\u2192\u9690\u85cf\u5c42（64\u795e\u7ecf\u5143）\u2192\u8f93\u51fa\u5c42（\u5bf9\u5e94\u7c7b\u522b\u6570），\u6fc0\u6d3b\u51fd\u6570\u7528ReLU，\u4f18\u5316\u5668\u4e3aAdam（lr=0.001），\u8bad\u7ec31500\u8f6e\u3002<\/p>\n<\/li>\n
- \n
1D-CNN（\u4e00\u7ef4\u5377\u79ef\u795e\u7ecf\u7f51\u7edc）：\u901a\u8fc7\u5377\u79ef\u6838\u6355\u6349\u5c40\u90e8\u7279\u5f81\u5173\u8054，\u9002\u914d\u77ff\u7269\u6210\u5206\u7684\u5e8f\u5217\u7279\u5f81\u3002\u7ed3\u6784：3\u5c42\u5377\u79ef（16\u219232\u219264\u901a\u9053）+\u5168\u8fde\u63a5\u5c42，\u8f93\u5165\u7ef4\u5ea6\u9002\u914d\u7279\u5f81\u6570，\u81ea\u52a8\u9002\u914dGPU\/CPU\u52a0\u901f\u8bad\u7ec3\u3002\u65e0\u9700\u50cfXGBoost\u90a3\u6837\u4e25\u683c\u5904\u7406\u6807\u7b7e，\u4f46\u4fdd\u7559\u8f7b\u91cf\u6620\u5c04\u903b\u8f91，\u907f\u514d\u795e\u7ecf\u5143\u6d6a\u8d39\u3002<\/p>\n<\/li>\n<\/ul>\n
2.4 \u7ed3\u679c\u5b58\u50a8\u8bbe\u8ba1<\/h5>\n
\u4e3a\u65b9\u4fbf\u540e\u7eed\u5bf9\u6bd4\u5206\u6790，\u5c06\u6240\u6709\u6a21\u578b\u7684\u8bc4\u4f30\u7ed3\u679c（\u51c6\u786e\u7387\u3001\u7cbe\u786e\u7387\u3001\u53ec\u56de\u7387\u3001F1\u5206\u6570）\u5b58\u5165JSON\u6587\u4ef6，\u91c7\u7528\u201c\u6a21\u578b\u2192\u586b\u5145\u65b9\u6cd5\u2192\u6307\u6807\u201d\u7684\u5d4c\u5957\u7ed3\u6784，\u786e\u4fdd\u6570\u636e\u53ef\u8ffd\u6eaf\u3001\u6613\u8bfb\u53d6\u3002<\/p>\n
\u4ee3\u7801：<\/p>\n
# import pandas as pd
\n# from sklearn.linear_model import LogisticRegression
\n# from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score
\n#
\n# # ———————- 1. \u914d\u7f6e\u53c2\u6570\u4e0e\u521d\u59cb\u5316 ———————-
\n# # \u5b9a\u4e49\u6240\u6709\u586b\u5145\u65b9\u6cd5（\u4e0e\u6587\u4ef6\u540d\u4e2d\u7684\u540e\u7f00\u5bf9\u5e94）
\n# fill_methods = ["\u4e2d\u4f4d\u6570", "\u4f17\u6570", "\u5220\u9664\u7a7a\u6570\u636e\u884c", "\u5747\u503c", "\u7ebf\u6027\u56de\u5f52", "\u968f\u673a\u68ee\u6797"]
\n# # \u521d\u59cb\u5316\u7ed3\u679c\u5b57\u5178
\n# lr_result_data = {}
\n# # \u903b\u8f91\u56de\u5f52\u6a21\u578b\u53c2\u6570（\u53ef\u6839\u636e\u9700\u8981\u8c03\u6574）
\n# logreg_params = {
\n# "max_iter": 1000, # \u589e\u52a0\u8fed\u4ee3\u6b21\u6570\u786e\u4fdd\u6536\u655b
\n# "random_state": 42,
\n# "n_jobs": 1
\n# }
\n#
\n# # ———————- 2. \u904d\u5386\u6240\u6709\u586b\u5145\u65b9\u6cd5，\u8bad\u7ec3+\u8bc4\u4f30 ———————-
\n# for method in fill_methods:
\n# # 1. \u8bfb\u53d6\u5f53\u524d\u586b\u5145\u65b9\u6cd5\u5bf9\u5e94\u7684\u8bad\u7ec3\u96c6\u548c\u6d4b\u8bd5\u96c6
\n# train_path = f"temp_data\/\u8bad\u7ec3\u6570\u636e\u96c6[{method}].xlsx"
\n# test_path = f"temp_data\/\u6d4b\u8bd5\u6570\u636e\u96c6[{method}].xlsx"
\n#
\n# try:
\n# train_df = pd.read_excel(train_path)
\n# test_df = pd.read_excel(test_path)
\n# except FileNotFoundError:
\n# print(f"\u8b66\u544a：\u672a\u627e\u5230 {method} \u5bf9\u5e94\u7684\u6570\u636e\u96c6\u6587\u4ef6，\u8df3\u8fc7\u8be5\u65b9\u6cd5")
\n# continue
\n#
\n# # 2. \u5206\u79bb\u7279\u5f81（X）\u548c\u6807\u7b7e（y，\u77ff\u7269\u7c7b\u578b）
\n# X_train = train_df.drop("\u77ff\u7269\u7c7b\u578b", axis=1)
\n# y_train = train_df["\u77ff\u7269\u7c7b\u578b"]
\n# X_test = test_df.drop("\u77ff\u7269\u7c7b\u578b", axis=1)
\n# y_test = test_df["\u77ff\u7269\u7c7b\u578b"]
\n#
\n# # 3. \u521d\u59cb\u5316\u5e76\u8bad\u7ec3\u903b\u8f91\u56de\u5f52\u6a21\u578b
\n# logreg = LogisticRegression(**logreg_params)
\n# logreg.fit(X_train, y_train)
\n#
\n# # 4. \u9884\u6d4b\u5e76\u8ba1\u7b97\u8bc4\u4f30\u6307\u6807
\n# y_pred = logreg.predict(X_test)
\n# metrics = {
\n# "\u51c6\u786e\u7387(Accuracy)": accuracy_score(y_test, y_pred),
\n# "\u7cbe\u786e\u7387(Precision)": precision_score(y_test, y_pred, average="weighted"),
\n# "\u53ec\u56de\u7387(Recall)": recall_score(y_test, y_pred, average="weighted"),
\n# "F1\u5206\u6570(F1-Score)": f1_score(y_test, y_pred, average="weighted")
\n# }
\n#
\n# # 5. \u5c06\u7ed3\u679c\u5b58\u5165\u5b57\u5178
\n# lr_result_data[method] = metrics
\n# print(f"\u2705 \u5b8c\u6210 {method} \u586b\u5145\u6570\u636e\u7684\u903b\u8f91\u56de\u5f52\u8bad\u7ec3\u4e0e\u8bc4\u4f30")
\n#
\n# # ———————- 3. \u6253\u5370\u7ed3\u679c\u5b57\u5178（\u53ef\u9009） ———————-
\n# print("\\\\n📊 \u6240\u6709\u586b\u5145\u65b9\u6cd5\u7684\u903b\u8f91\u56de\u5f52\u5206\u7c7b\u7ed3\u679c：")
\n# for method, metrics in lr_result_data.items():
\n# print(f"\\\\n— {method} —")
\n# for metric_name, value in metrics.items():
\n# print(f"{metric_name}: {value:.4f}")
\n#
\n#
\n# import json
\n#
\n# # ———————- 1. \u521d\u59cb\u5316\u603b\u7ed3\u679c\u5b57\u5178（\u7528\u4e8e\u6240\u6709\u6a21\u578b\u7684\u5bf9\u6bd4） ———————-
\n# all_model_results = {}
\n#
\n# # \u793a\u4f8b：\u8ffd\u52a0\u968f\u673a\u68ee\u6797\u5206\u7c7b\u7684\u7ed3\u679c
\n# all_model_results["\u903b\u8f91\u56de\u5f52"] = lr_result_data # rf_result_data \u662f\u968f\u673a\u68ee\u6797\u7684\u7ed3\u679c\u5b57\u5178
\n#
\n# # \u518d\u6b21\u4fdd\u5b58\u5230 JSON（\u8986\u76d6\u539f\u6587\u4ef6，\u6216\u7528 'a' \u6a21\u5f0f\u8ffd\u52a0）
\n# with open(r'temp_data\/\u6240\u6709\u6a21\u578b\u5206\u7c7b\u7ed3\u679c.json', 'w', encoding='utf-8') as file:
\n# json.dump(all_model_results, file, ensure_ascii=False, indent=4)
\n#
\n# import pandas as pd
\n# from sklearn.svm import SVC
\n# from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score
\n# import json
\n#
\n# # ———————- 1. \u914d\u7f6e\u53c2\u6570\u4e0e\u521d\u59cb\u5316 ———————-
\n# # \u5b9a\u4e49\u6240\u6709\u586b\u5145\u65b9\u6cd5（\u4e0e\u6587\u4ef6\u540d\u5bf9\u5e94）
\n# fill_methods = ["\u4e2d\u4f4d\u6570", "\u4f17\u6570", "\u5220\u9664\u7a7a\u6570\u636e\u884c", "\u5747\u503c", "\u7ebf\u6027\u56de\u5f52", "\u968f\u673a\u68ee\u6797"]
\n# # \u521d\u59cb\u5316 SVM \u7ed3\u679c\u5b57\u5178
\n# svm_result_data = {}
\n# # \u4f60\u7684 SVM \u53c2\u6570（\u5b8c\u5168\u6309\u7167\u56fe\u7247\u914d\u7f6e）
\n# # \u4fee\u6539\u540e\u7684 SVM \u53c2\u6570（\u5927\u5e45\u63d0\u5347\u901f\u5ea6）
\n# svm_params = {
\n# "C": 1,
\n# "coef0": 0.1,
\n# "degree": 2, # \u964d\u4f4e\u591a\u9879\u5f0f\u9636\u6570（\u4ece4\u21922），\u51cf\u5c11\u8ba1\u7b97\u91cf
\n# "gamma": "scale", # \u81ea\u52a8\u7f29\u653egamma，\u907f\u514d\u624b\u52a8\u8bbe\u7f6e\u5bfc\u81f4\u7684\u6570\u503c\u95ee\u9898
\n# "kernel": "poly",
\n# "probability": False, # \u82e5\u4e0d\u9700\u8981\u6982\u7387\u8f93\u51fa，\u8bbe\u4e3aFalse（\u5927\u5e45\u63d0\u901f）
\n# "random_state": 100,
\n# "cache_size": 2000 # \u589e\u52a0\u7f13\u5b58\u5927\u5c0f（\u5355\u4f4dMB），\u52a0\u901f\u8ba1\u7b97
\n# }
\n#
\n# # ———————- 2. \u904d\u5386\u6240\u6709\u586b\u5145\u65b9\u6cd5，\u8bad\u7ec3+\u8bc4\u4f30 SVM ———————-
\n# for method in fill_methods:
\n# # \u8bfb\u53d6\u5f53\u524d\u586b\u5145\u65b9\u6cd5\u7684\u8bad\u7ec3\u96c6\u548c\u6d4b\u8bd5\u96c6
\n# train_path = f"temp_data\/\u8bad\u7ec3\u6570\u636e\u96c6[{method}].xlsx"
\n# test_path = f"temp_data\/\u6d4b\u8bd5\u6570\u636e\u96c6[{method}].xlsx"
\n#
\n# try:
\n# train_df = pd.read_excel(train_path)
\n# test_df = pd.read_excel(test_path)
\n# except FileNotFoundError:
\n# print(f"\u8b66\u544a：\u672a\u627e\u5230 {method} \u5bf9\u5e94\u7684\u6570\u636e\u96c6，\u8df3\u8fc7\u8be5\u65b9\u6cd5")
\n# continue
\n#
\n# # \u5206\u79bb\u7279\u5f81（X）\u548c\u6807\u7b7e（y，\u77ff\u7269\u7c7b\u578b）
\n# X_train = train_df.drop("\u77ff\u7269\u7c7b\u578b", axis=1)
\n# y_train = train_df["\u77ff\u7269\u7c7b\u578b"]
\n# X_test = test_df.drop("\u77ff\u7269\u7c7b\u578b", axis=1)
\n# y_test = test_df["\u77ff\u7269\u7c7b\u578b"]
\n#
\n# # \u521d\u59cb\u5316\u5e76\u8bad\u7ec3 SVM \u6a21\u578b
\n# svm = SVC(**svm_params)
\n# svm.fit(X_train, y_train)
\n#
\n# # \u9884\u6d4b\u5e76\u8ba1\u7b97\u8bc4\u4f30\u6307\u6807
\n# y_pred = svm.predict(X_test)
\n# metrics = {
\n# "\u51c6\u786e\u7387(Accuracy)": accuracy_score(y_test, y_pred),
\n# "\u7cbe\u786e\u7387(Precision)": precision_score(y_test, y_pred, average="weighted"),
\n# "\u53ec\u56de\u7387(Recall)": recall_score(y_test, y_pred, average="weighted"),
\n# "F1\u5206\u6570(F1-Score)": f1_score(y_test, y_pred, average="weighted")
\n# }
\n#
\n# # \u5c06\u7ed3\u679c\u5b58\u5165 SVM \u7ed3\u679c\u5b57\u5178
\n# svm_result_data[method] = metrics
\n# print(f"\u2705 \u5b8c\u6210 {method} \u586b\u5145\u6570\u636e\u7684 SVM \u8bad\u7ec3\u4e0e\u8bc4\u4f30")
\n#
\n# # ———————- 3. \u8ffd\u52a0\u5230\u603b\u7ed3\u679c\u5b57\u5178\u5e76\u4fdd\u5b58 ———————-
\n# # \u8bfb\u53d6\u4e4b\u524d\u7684\u603b\u7ed3\u679c（\u5982\u679c\u5b58\u5728）
\n# try:
\n# with open(r'temp_data\/\u6240\u6709\u6a21\u578b\u5206\u7c7b\u7ed3\u679c.json', 'r', encoding='utf-8') as file:
\n# all_model_results = json.load(file)
\n# except FileNotFoundError:
\n# all_model_results = {}
\n#
\n# # \u8ffd\u52a0 SVM \u7ed3\u679c\u5230\u603b\u5b57\u5178
\n# all_model_results["\u652f\u6301\u5411\u91cf\u673a（Poly\u6838）"] = svm_result_data
\n#
\n# # \u4fdd\u5b58\u66f4\u65b0\u540e\u7684\u603b\u7ed3\u679c
\n# with open(r'temp_data\/\u6240\u6709\u6a21\u578b\u5206\u7c7b\u7ed3\u679c.json', 'w', encoding='utf-8') as file:
\n# json.dump(all_model_results, file, ensure_ascii=False, indent=4)
\n#
\n# print("\\\\n📊 SVM \u5206\u7c7b\u7ed3\u679c\u5df2\u8ffd\u52a0\u5230 temp_data\/\u6240\u6709\u6a21\u578b\u5206\u7c7b\u7ed3\u679c.json")
\n#<\/p>\n# import pandas as pd
\n# from sklearn.ensemble import RandomForestClassifier
\n# from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score
\n# import json
\n#
\n# # ———————- 1. \u914d\u7f6e\u53c2\u6570\u4e0e\u521d\u59cb\u5316 ———————-
\n# # \u5b9a\u4e49\u6240\u6709\u586b\u5145\u65b9\u6cd5（\u4e0e\u6587\u4ef6\u540d\u5bf9\u5e94）
\n# fill_methods = ["\u4e2d\u4f4d\u6570", "\u4f17\u6570", "\u5220\u9664\u7a7a\u6570\u636e\u884c", "\u5747\u503c", "\u7ebf\u6027\u56de\u5f52", "\u968f\u673a\u68ee\u6797"]
\n# # \u521d\u59cb\u5316\u968f\u673a\u68ee\u6797\u7ed3\u679c\u5b57\u5178
\n# rf_result_data = {}
\n# # \u4f60\u7684\u968f\u673a\u68ee\u6797\u53c2\u6570（\u5b8c\u5168\u6309\u7167\u56fe\u7247\u914d\u7f6e）
\n# rf_params = {
\n# "bootstrap": False,
\n# "max_depth": 20,
\n# "max_features": "log2",
\n# "min_samples_leaf": 1,
\n# "min_samples_split": 2,
\n# "n_estimators": 50,
\n# "random_state": 487,
\n# "n_jobs": -1 # \u591a\u7ebf\u7a0b\u52a0\u901f\u8bad\u7ec3（\u53ef\u9009，\u6839\u636e\u786c\u4ef6\u8c03\u6574）
\n# }
\n#
\n# # ———————- 2. \u904d\u5386\u6240\u6709\u586b\u5145\u65b9\u6cd5，\u8bad\u7ec3+\u8bc4\u4f30\u968f\u673a\u68ee\u6797 ———————-
\n# for method in fill_methods:
\n# # \u8bfb\u53d6\u5f53\u524d\u586b\u5145\u65b9\u6cd5\u7684\u8bad\u7ec3\u96c6\u548c\u6d4b\u8bd5\u96c6
\n# train_path = f"temp_data\/\u8bad\u7ec3\u6570\u636e\u96c6[{method}].xlsx"
\n# test_path = f"temp_data\/\u6d4b\u8bd5\u6570\u636e\u96c6[{method}].xlsx"
\n#
\n# try:
\n# train_df = pd.read_excel(train_path)
\n# test_df = pd.read_excel(test_path)
\n# except FileNotFoundError:
\n# print(f"\u8b66\u544a：\u672a\u627e\u5230 {method} \u5bf9\u5e94\u7684\u6570\u636e\u96c6，\u8df3\u8fc7\u8be5\u65b9\u6cd5")
\n# continue
\n#
\n# # \u5206\u79bb\u7279\u5f81（X）\u548c\u6807\u7b7e（y，\u77ff\u7269\u7c7b\u578b）
\n# X_train = train_df.drop("\u77ff\u7269\u7c7b\u578b", axis=1)
\n# y_train = train_df["\u77ff\u7269\u7c7b\u578b"]
\n# X_test = test_df.drop("\u77ff\u7269\u7c7b\u578b", axis=1)
\n# y_test = test_df["\u77ff\u7269\u7c7b\u578b"]
\n#
\n# # \u521d\u59cb\u5316\u5e76\u8bad\u7ec3\u968f\u673a\u68ee\u6797\u6a21\u578b
\n# rf = RandomForestClassifier(**rf_params)
\n# rf.fit(X_train, y_train)
\n#
\n# # \u9884\u6d4b\u5e76\u8ba1\u7b97\u8bc4\u4f30\u6307\u6807
\n# y_pred = rf.predict(X_test)
\n# metrics = {
\n# "\u51c6\u786e\u7387(Accuracy)": accuracy_score(y_test, y_pred),
\n# "\u7cbe\u786e\u7387(Precision)": precision_score(y_test, y_pred, average="weighted"),
\n# "\u53ec\u56de\u7387(Recall)": recall_score(y_test, y_pred, average="weighted"),
\n# "F1\u5206\u6570(F1-Score)": f1_score(y_test, y_pred, average="weighted")
\n# }
\n#
\n# # \u5c06\u7ed3\u679c\u5b58\u5165\u968f\u673a\u68ee\u6797\u7ed3\u679c\u5b57\u5178
\n# rf_result_data[method] = metrics
\n# print(f"\u2705 \u5b8c\u6210 {method} \u586b\u5145\u6570\u636e\u7684\u968f\u673a\u68ee\u6797\u8bad\u7ec3\u4e0e\u8bc4\u4f30")
\n#
\n# # ———————- 3. \u8ffd\u52a0\u5230\u603b\u7ed3\u679c\u5b57\u5178\u5e76\u4fdd\u5b58 ———————-
\n# # \u8bfb\u53d6\u4e4b\u524d\u7684\u603b\u7ed3\u679c（\u5982\u679c\u5b58\u5728）
\n# try:
\n# with open(r'temp_data\/\u6240\u6709\u6a21\u578b\u5206\u7c7b\u7ed3\u679c.json', 'r', encoding='utf-8') as file:
\n# all_model_results = json.load(file)
\n# except FileNotFoundError:
\n# all_model_results = {}
\n#
\n# # \u8ffd\u52a0\u968f\u673a\u68ee\u6797\u7ed3\u679c\u5230\u603b\u5b57\u5178
\n# all_model_results["\u968f\u673a\u68ee\u6797\u5206\u7c7b"] = rf_result_data
\n#
\n# # \u4fdd\u5b58\u66f4\u65b0\u540e\u7684\u603b\u7ed3\u679c
\n# with open(r'temp_data\/\u6240\u6709\u6a21\u578b\u5206\u7c7b\u7ed3\u679c.json', 'w', encoding='utf-8') as file:
\n# json.dump(all_model_results, file, ensure_ascii=False, indent=4)
\n#
\n# print("\\\\n📊 \u968f\u673a\u68ee\u6797\u5206\u7c7b\u7ed3\u679c\u5df2\u8ffd\u52a0\u5230 temp_data\/\u6240\u6709\u6a21\u578b\u5206\u7c7b\u7ed3\u679c.json")<\/p>\n# import pandas as pd
\n# from sklearn.ensemble import AdaBoostClassifier
\n# from sklearn.tree import DecisionTreeClassifier
\n# from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score
\n# import json
\n#
\n# # ———————- 1. \u914d\u7f6e\u53c2\u6570\u4e0e\u521d\u59cb\u5316 ———————-
\n# fill_methods = ["\u4e2d\u4f4d\u6570", "\u4f17\u6570", "\u5220\u9664\u7a7a\u6570\u636e\u884c", "\u5747\u503c", "\u7ebf\u6027\u56de\u5f52", "\u968f\u673a\u68ee\u6797"]
\n# abf_result_data = {}
\n#
\n# # \u5b8c\u5168\u6309\u7167\u56fe\u7247\u914d\u7f6e AdaBoost \u53c2\u6570
\n# abf_params = {
\n# "algorithm": "SAMME",
\n# "base_estimator": DecisionTreeClassifier(max_depth=2),
\n# "n_estimators": 200,
\n# "learning_rate": 1.0,
\n# "random_state": 0
\n# }
\n#
\n# # ———————- 2. \u904d\u5386\u6240\u6709\u586b\u5145\u65b9\u6cd5，\u8bad\u7ec3+\u8bc4\u4f30 AdaBoost ———————-
\n# for method in fill_methods:
\n# train_path = f"temp_data\/\u8bad\u7ec3\u6570\u636e\u96c6[{method}].xlsx"
\n# test_path = f"temp_data\/\u6d4b\u8bd5\u6570\u636e\u96c6[{method}].xlsx"
\n#
\n# try:
\n# train_df = pd.read_excel(train_path)
\n# test_df = pd.read_excel(test_path)
\n# except FileNotFoundError:
\n# print(f"\u8b66\u544a：\u672a\u627e\u5230 {method} \u5bf9\u5e94\u7684\u6570\u636e\u96c6，\u8df3\u8fc7\u8be5\u65b9\u6cd5")
\n# continue
\n#
\n# # \u5206\u79bb\u7279\u5f81\u4e0e\u6807\u7b7e
\n# X_train = train_df.drop("\u77ff\u7269\u7c7b\u578b", axis=1)
\n# y_train = train_df["\u77ff\u7269\u7c7b\u578b"]
\n# X_test = test_df.drop("\u77ff\u7269\u7c7b\u578b", axis=1)
\n# y_test = test_df["\u77ff\u7269\u7c7b\u578b"]
\n#
\n# # \u521d\u59cb\u5316\u5e76\u8bad\u7ec3 AdaBoost
\n# abf = AdaBoostClassifier(**abf_params)
\n# abf.fit(X_train, y_train)
\n#
\n# # \u9884\u6d4b\u4e0e\u8bc4\u4f30
\n# y_pred = abf.predict(X_test)
\n# metrics = {
\n# "\u51c6\u786e\u7387(Accuracy)": accuracy_score(y_test, y_pred),
\n# "\u7cbe\u786e\u7387(Precision)": precision_score(y_test, y_pred, average="weighted"),
\n# "\u53ec\u56de\u7387(Recall)": recall_score(y_test, y_pred, average="weighted"),
\n# "F1\u5206\u6570(F1-Score)": f1_score(y_test, y_pred, average="weighted")
\n# }
\n#
\n# abf_result_data[method] = metrics
\n# print(f"\u2705 \u5b8c\u6210 {method} \u586b\u5145\u6570\u636e\u7684 AdaBoost \u8bad\u7ec3\u4e0e\u8bc4\u4f30")
\n#
\n# # ———————- 3. \u8ffd\u52a0\u5230\u603b\u7ed3\u679c\u5b57\u5178\u5e76\u4fdd\u5b58 ———————-
\n# try:
\n# with open(r'temp_data\/\u6240\u6709\u6a21\u578b\u5206\u7c7b\u7ed3\u679c.json', 'r', encoding='utf-8') as file:
\n# all_model_results = json.load(file)
\n# except FileNotFoundError:
\n# all_model_results = {}
\n#
\n# all_model_results["AdaBoost\u5206\u7c7b"] = abf_result_data
\n#
\n# with open(r'temp_data\/\u6240\u6709\u6a21\u578b\u5206\u7c7b\u7ed3\u679c.json', 'w', encoding='utf-8') as file:
\n# json.dump(all_model_results, file, ensure_ascii=False, indent=4)
\n#
\n# print("\\\\n📊 AdaBoost \u5206\u7c7b\u7ed3\u679c\u5df2\u8ffd\u52a0\u5230 temp_data\/\u6240\u6709\u6a21\u578b\u5206\u7c7b\u7ed3\u679c.json")<\/p>\n#
\n# import pandas as pd
\n# import xgboost as xgb
\n# from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score
\n# import json
\n#
\n# # ———————- 1. \u914d\u7f6e\u53c2\u6570\u4e0e\u521d\u59cb\u5316 ———————-
\n# # \u5b9a\u4e49\u6240\u6709\u586b\u5145\u65b9\u6cd5（\u4e0e\u6587\u4ef6\u540d\u5bf9\u5e94）
\n# fill_methods = ["\u4e2d\u4f4d\u6570", "\u4f17\u6570", "\u5220\u9664\u7a7a\u6570\u636e\u884c", "\u5747\u503c", "\u7ebf\u6027\u56de\u5f52", "\u968f\u673a\u68ee\u6797"]
\n# # \u521d\u59cb\u5316XGBoost\u7ed3\u679c\u5b57\u5178
\n# xgb_result_data = {}
\n#
\n# # XGBoost\u57fa\u7840\u53c2\u6570（num_class\u52a8\u6001\u8bbe\u7f6e，objective\u517c\u5bb9\u4e0d\u8fde\u7eed\u6807\u7b7e）
\n# xgb_base_params = {
\n# "learning_rate": 0.05, # \u5b66\u4e60\u7387
\n# "n_estimators": 200, # \u51b3\u7b56\u6811\u6570\u91cf
\n# "max_depth": 7, # \u6811\u7684\u6700\u5927\u6df1\u5ea6
\n# "min_child_weight": 1, # \u53f6\u5b50\u8282\u70b9\u6700\u5c0f\u6837\u672c\u6743\u91cd\u548c
\n# "gamma": 0, # \u8282\u70b9\u5206\u88c2\u6240\u9700\u6700\u5c0f\u635f\u5931\u4e0b\u964d\u503c
\n# "subsample": 0.6, # \u8bad\u7ec3\u6837\u672c\u5b50\u6837\u672c\u6bd4\u4f8b
\n# "colsample_bytree": 0.8, # \u6bcf\u68f5\u6811\u968f\u673a\u91c7\u6837\u5217\u5360\u6bd4
\n# "objective": "multi:softprob", # \u66ff\u6362\u4e3asoftprob，\u517c\u5bb9\u4e0d\u8fde\u7eed\u6807\u7b7e
\n# "seed": 0 # \u968f\u673a\u6570\u79cd\u5b50
\n# }
\n#
\n# # ———————- 2. \u904d\u5386\u6240\u6709\u586b\u5145\u65b9\u6cd5，\u8bad\u7ec3+\u8bc4\u4f30 XGBoost ———————-
\n# for method in fill_methods:
\n# # \u8bfb\u53d6\u5f53\u524d\u586b\u5145\u65b9\u6cd5\u7684\u8bad\u7ec3\u96c6\u548c\u6d4b\u8bd5\u96c6
\n# train_path = f"temp_data\/\u8bad\u7ec3\u6570\u636e\u96c6[{method}].xlsx"
\n# test_path = f"temp_data\/\u6d4b\u8bd5\u6570\u636e\u96c6[{method}].xlsx"
\n#
\n# try:
\n# train_df = pd.read_excel(train_path)
\n# test_df = pd.read_excel(test_path)
\n# except FileNotFoundError:
\n# print(f"\u26a0\ufe0f \u8b66\u544a：\u672a\u627e\u5230 {method} \u5bf9\u5e94\u7684\u6570\u636e\u96c6，\u8df3\u8fc7\u8be5\u65b9\u6cd5")
\n# continue
\n#
\n# # \u5206\u79bb\u7279\u5f81\u4e0e\u539f\u59cb\u6807\u7b7e
\n# X_train = train_df.drop("\u77ff\u7269\u7c7b\u578b", axis=1)
\n# y_train_original = train_df["\u77ff\u7269\u7c7b\u578b"]
\n# X_test = test_df.drop("\u77ff\u7269\u7c7b\u578b", axis=1)
\n# y_test_original = test_df["\u77ff\u7269\u7c7b\u578b"]
\n#
\n# # ———————- \u6838\u5fc3\u4fee\u590d：\u6807\u7b7e\u8fde\u7eed\u6027\u5904\u7406 ———————-
\n# # 1. \u83b7\u53d6\u8bad\u7ec3\u96c6\u552f\u4e00\u6807\u7b7e\u5e76\u6392\u5e8f
\n# unique_train_classes = sorted(y_train_original.unique())
\n# print(f"\\\\n📌 {method} \u586b\u5145\u6570\u636e\u96c6 – \u539f\u59cb\u6807\u7b7e\u7c7b\u522b：{unique_train_classes}")
\n#
\n# # 2. \u6784\u5efa\u6807\u7b7e\u6620\u5c04（\u5c06\u539f\u59cb\u6807\u7b7e\u6620\u5c04\u4e3a\u4ece0\u5f00\u59cb\u7684\u8fde\u7eed\u6574\u6570）
\n# class_mapping = {old_label: new_label for new_label, old_label in enumerate(unique_train_classes)}
\n# print(f"🔄 \u6807\u7b7e\u6620\u5c04\u89c4\u5219：{class_mapping}")
\n#
\n# # 3. \u5e94\u7528\u6620\u5c04，\u786e\u4fdd\u6807\u7b7e\u8fde\u7eed
\n# y_train = y_train_original.map(class_mapping)
\n# y_test = y_test_original.map(class_mapping)
\n#
\n# # 4. \u52a8\u6001\u8bbe\u7f6e\u7c7b\u522b\u6570（\u907f\u514d\u56fa\u5b9anum_class\u5bfc\u81f4\u4e0d\u5339\u914d）
\n# xgb_params = xgb_base_params.copy()
\n# xgb_params["num_class"] = len(unique_train_classes)
\n#
\n# # ———————- \u6a21\u578b\u8bad\u7ec3\u4e0e\u8bc4\u4f30 ———————-
\n# # \u521d\u59cb\u5316\u5e76\u8bad\u7ec3XGBoost
\n# xgb_model = xgb.XGBClassifier(**xgb_params)
\n# xgb_model.fit(X_train, y_train)
\n#
\n# # \u9884\u6d4b（\u8fd4\u56de\u7684\u662f\u6620\u5c04\u540e\u7684\u8fde\u7eed\u6807\u7b7e）
\n# y_pred = xgb_model.predict(X_test)
\n#
\n# # \u8ba1\u7b97\u8bc4\u4f30\u6307\u6807（weighted\u9002\u914d\u591a\u5206\u7c7b）
\n# metrics = {
\n# "\u51c6\u786e\u7387(Accuracy)": round(accuracy_score(y_test, y_pred), 4),
\n# "\u7cbe\u786e\u7387(Precision)": round(precision_score(y_test, y_pred, average="weighted", zero_division=0), 4),
\n# "\u53ec\u56de\u7387(Recall)": round(recall_score(y_test, y_pred, average="weighted", zero_division=0), 4),
\n# "F1\u5206\u6570(F1-Score)": round(f1_score(y_test, y_pred, average="weighted", zero_division=0), 4)
\n# }
\n#
\n# # \u4fdd\u5b58\u7ed3\u679c
\n# xgb_result_data[method] = metrics
\n# print(f"\u2705 \u5b8c\u6210 {method} \u586b\u5145\u6570\u636e\u7684 XGBoost \u8bad\u7ec3 | \u51c6\u786e\u7387：{metrics['\u51c6\u786e\u7387(Accuracy)']}")
\n#
\n# # ———————- 3. \u8ffd\u52a0\u5230\u603b\u7ed3\u679c\u5b57\u5178\u5e76\u4fdd\u5b58 ———————-
\n# try:
\n# # \u8bfb\u53d6\u5df2\u6709\u7684\u603b\u7ed3\u679c\u6587\u4ef6
\n# with open(r'temp_data\/\u6240\u6709\u6a21\u578b\u5206\u7c7b\u7ed3\u679c.json', 'r', encoding='utf-8') as file:
\n# all_model_results = json.load(file)
\n# except FileNotFoundError:
\n# # \u82e5\u6587\u4ef6\u4e0d\u5b58\u5728，\u521d\u59cb\u5316\u7a7a\u5b57\u5178
\n# all_model_results = {}
\n#
\n# # \u8ffd\u52a0XGBoost\u7ed3\u679c
\n# all_model_results["XGBoost\u5206\u7c7b"] = xgb_result_data
\n#
\n# # \u4fdd\u5b58\u66f4\u65b0\u540e\u7684\u603b\u7ed3\u679c
\n# with open(r'temp_data\/\u6240\u6709\u6a21\u578b\u5206\u7c7b\u7ed3\u679c.json', 'w', encoding='utf-8') as file:
\n# json.dump(all_model_results, file, ensure_ascii=False, indent=4)
\n#
\n# print("\\\\n📊 XGBoost \u5206\u7c7b\u7ed3\u679c\u5df2\u8ffd\u52a0\u5230 temp_data\/\u6240\u6709\u6a21\u578b\u5206\u7c7b\u7ed3\u679c.json")
\n#
\n#<\/p>\nimport pandas as pd
\nimport torch
\nimport torch.nn as nn
\nimport torch.optim as optim
\nimport numpy as np
\nfrom sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score
\nimport json<\/p>\n# ———————- 1. \u5b9a\u4e49\u6a21\u578b\u7ed3\u6784 ———————-
\n# \u591a\u5c42\u611f\u77e5\u673a（MLP）
\nclass MLP(nn.Module):
\n def __init__(self, input_dim, hidden_dim, num_classes):
\n super(MLP, self).__init__()
\n self.fc1 = nn.Linear(input_dim, hidden_dim)
\n self.fc2 = nn.Linear(hidden_dim, hidden_dim * 2)
\n self.fc3 = nn.Linear(hidden_dim * 2, num_classes)<\/p>\ndef forward(self, x):
\n x = torch.relu(self.fc1(x))
\n x = torch.relu(self.fc2(x))
\n x = self.fc3(x)
\n return x<\/p>\n# \u4e00\u7ef4\u5377\u79ef\u795e\u7ecf\u7f51\u7edc（CNN）
\nclass CNN1D(nn.Module):
\n def __init__(self, input_dim, hidden_size, num_classes):
\n super(CNN1D, self).__init__()
\n self.conv1 = nn.Conv1d(in_channels=1, out_channels=16, kernel_size=3, padding=1)
\n self.conv2 = nn.Conv1d(in_channels=16, out_channels=32, kernel_size=3, padding=1)
\n self.conv3 = nn.Conv1d(in_channels=32, out_channels=64, kernel_size=3, padding=1)
\n self.relu = nn.ReLU()
\n self.fc = nn.Linear(64 * input_dim, num_classes) # \u8f93\u5165\u7ef4\u5ea6\u9002\u914d<\/p>\ndef forward(self, x):
\n x = x.unsqueeze(1) # \u589e\u52a0\u901a\u9053\u7ef4\u5ea6（batch_size, 1, input_dim）
\n x = self.relu(self.conv1(x))
\n x = self.relu(self.conv2(x))
\n x = self.relu(self.conv3(x))
\n x = x.view(x.size(0), -1) # \u5c55\u5e73
\n x = self.fc(x)
\n return x<\/p>\n# ———————- 2. \u8bad\u7ec3\u4e0e\u8bc4\u4f30\u51fd\u6570 ———————-
\ndef train_evaluate_model(model, X_train, y_train, X_test, y_test, criterion, optimizer, epochs=1500,
\n print_interval=100):
\n for epoch in range(epochs):
\n # \u8bad\u7ec3\u6a21\u5f0f
\n model.train()
\n optimizer.zero_grad()
\n outputs = model(X_train)
\n loss = criterion(outputs, y_train)
\n loss.backward()
\n optimizer.step()<\/p>\n# \u6bcfprint_interval\u8f6e\u6253\u5370\u4e00\u6b21
\n if (epoch + 1) % print_interval == 0:
\n model.eval()
\n with torch.no_grad():
\n test_pred = model(X_test).argmax(dim=1)
\n test_acc = accuracy_score(y_test.cpu().numpy(), test_pred.cpu().numpy())
\n print(f"Epoch [{epoch + 1}\/{epochs}] | Loss: {loss.item():.4f} | Test Acc: {test_acc:.4f}")<\/p>\n# \u6700\u7ec8\u8bc4\u4f30
\n model.eval()
\n with torch.no_grad():
\n y_pred = model(X_test).argmax(dim=1)
\n # \u8f6c\u5230CPU\u8ba1\u7b97\u6307\u6807（\u517c\u5bb9GPU\u8bad\u7ec3）
\n y_test_np = y_test.cpu().numpy()
\n y_pred_np = y_pred.cpu().numpy()
\n metrics = {
\n "\u51c6\u786e\u7387(Accuracy)": round(accuracy_score(y_test_np, y_pred_np), 4),
\n "\u7cbe\u786e\u7387(Precision)": round(precision_score(y_test_np, y_pred_np, average="weighted", zero_division=0), 4),
\n "\u53ec\u56de\u7387(Recall)": round(recall_score(y_test_np, y_pred_np, average="weighted", zero_division=0), 4),
\n "F1\u5206\u6570(F1-Score)": round(f1_score(y_test_np, y_pred_np, average="weighted", zero_division=0), 4)
\n }
\n return metrics<\/p>\n# ———————- 3. \u914d\u7f6e\u53c2\u6570\u4e0e\u521d\u59cb\u5316 ———————-
\nfill_methods = ["\u4e2d\u4f4d\u6570", "\u4f17\u6570", "\u5220\u9664\u7a7a\u6570\u636e\u884c", "\u5747\u503c", "\u7ebf\u6027\u56de\u5f52", "\u968f\u673a\u68ee\u6797"]
\nmlp_result_data = {}
\ncnn_result_data = {}
\n# \u81ea\u52a8\u9009\u62e9\u8bbe\u5907（GPU\/CPU）
\ndevice = torch.device("cuda" if torch.cuda.is_available() else "cpu")
\nprint(f"\u4f7f\u7528\u8bbe\u5907: {device}")<\/p>\n# ———————- 4. \u904d\u5386\u6240\u6709\u586b\u5145\u65b9\u6cd5\u8bad\u7ec3 ———————-
\nfor method in fill_methods:
\n train_path = f"temp_data\/\u8bad\u7ec3\u6570\u636e\u96c6[{method}].xlsx"
\n test_path = f"temp_data\/\u6d4b\u8bd5\u6570\u636e\u96c6[{method}].xlsx"<\/p>\ntry:
\n train_df = pd.read_excel(train_path)
\n test_df = pd.read_excel(test_path)
\n except FileNotFoundError:
\n print(f"\u26a0\ufe0f \u8b66\u544a：\u672a\u627e\u5230 {method} \u5bf9\u5e94\u7684\u6570\u636e\u96c6，\u8df3\u8fc7\u8be5\u65b9\u6cd5")
\n continue<\/p>\n# \u5206\u79bb\u7279\u5f81\u4e0e\u6807\u7b7e（\u539f\u59cb\u6570\u636e）
\n X_train_np = train_df.drop("\u77ff\u7269\u7c7b\u578b", axis=1).values.astype(np.float32)
\n y_train_original = train_df["\u77ff\u7269\u7c7b\u578b"].values
\n X_test_np = test_df.drop("\u77ff\u7269\u7c7b\u578b", axis=1).values.astype(np.float32)
\n y_test_original = test_df["\u77ff\u7269\u7c7b\u578b"].values<\/p>\n# \u2705 \u8f7b\u91cf\u6807\u7b7e\u5904\u7406（\u4ec5\u6838\u5fc3\u6620\u5c04，\u65e0\u5197\u4f59\u6253\u5370）
\n # \u6620\u5c04\u4e3a\u4ece0\u5f00\u59cb\u7684\u8fde\u7eed\u6574\u6570（\u907f\u514d\u795e\u7ecf\u5143\u6d6a\u8d39\/\u7ef4\u5ea6\u4e0d\u5339\u914d）
\n unique_classes = sorted(np.unique(y_train_original))
\n class_mapping = {old: new for new, old in enumerate(unique_classes)}
\n y_train_np = np.array([class_mapping[label] for label in y_train_original])
\n y_test_np = np.array([class_mapping[label] for label in y_test_original])
\n num_classes = len(unique_classes)
\n input_dim = X_train_np.shape[1]<\/p>\n# \u8f6c\u6362\u4e3aPyTorch\u5f20\u91cf（\u8f6c\u5230\u6307\u5b9a\u8bbe\u5907）
\n X_train = torch.tensor(X_train_np).to(device)
\n y_train = torch.tensor(y_train_np, dtype=torch.long).to(device)
\n X_test = torch.tensor(X_test_np).to(device)
\n y_test = torch.tensor(y_test_np, dtype=torch.long).to(device)<\/p>\n# ———————- \u8bad\u7ec3MLP ———————-
\n print(f"\\\\n=== \u8bad\u7ec3 {method} \u586b\u5145\u6570\u636e\u7684 MLP ===")
\n mlp = MLP(input_dim=input_dim, hidden_dim=32, num_classes=num_classes).to(device)
\n criterion = nn.CrossEntropyLoss()
\n optimizer = optim.Adam(mlp.parameters(), lr=0.001)
\n mlp_metrics = train_evaluate_model(mlp, X_train, y_train, X_test, y_test, criterion, optimizer)
\n mlp_result_data[method] = mlp_metrics<\/p>\n# ———————- \u8bad\u7ec3CNN ———————-
\n print(f"\\\\n=== \u8bad\u7ec3 {method} \u586b\u5145\u6570\u636e\u7684 CNN ===")
\n cnn = CNN1D(input_dim=input_dim, hidden_size=10, num_classes=num_classes).to(device)
\n optimizer = optim.Adam(cnn.parameters(), lr=0.001)
\n cnn_metrics = train_evaluate_model(cnn, X_train, y_train, X_test, y_test, criterion, optimizer)
\n cnn_result_data[method] = cnn_metrics<\/p>\n# ———————- 5. \u4fdd\u5b58\u7ed3\u679c ———————-
\ntry:
\n with open(r'temp_data\/\u6240\u6709\u6a21\u578b\u5206\u7c7b\u7ed3\u679c.json', 'r', encoding='utf-8') as file:
\n all_model_results = json.load(file)
\nexcept FileNotFoundError:
\n all_model_results = {}<\/p>\nall_model_results["\u795e\u7ecf\u7f51\u7edc（MLP）\u5206\u7c7b"] = mlp_result_data
\nall_model_results["\u5377\u79ef\u795e\u7ecf\u7f51\u7edc（CNN）\u5206\u7c7b"] = cnn_result_data<\/p>\nwith open(r'temp_data\/\u6240\u6709\u6a21\u578b\u5206\u7c7b\u7ed3\u679c.json', 'w', encoding='utf-8') as file:
\n json.dump(all_model_results, file, ensure_ascii=False, indent=4)<\/p>\nprint("\\\\n📊 \u795e\u7ecf\u7f51\u7edc\/CNN\u7ed3\u679c\u5df2\u4fdd\u5b58\u5230 temp_data\/\u6240\u6709\u6a21\u578b\u5206\u7c7b\u7ed3\u679c.json")<\/p>\n
\u9636\u6bb5\u4e09：\u7ed3\u679c\u5206\u6790\u4e0e\u6700\u4f18\u65b9\u6848\u7b5b\u9009<\/h4>\n
\u901a\u8fc7\u5bf9\u6bd442\u79cd\u65b9\u6848\u7684\u8bc4\u4f30\u6307\u6807，\u6700\u7ec8\u53d1\u73b0XGBoost+\u4e2d\u4f4d\u6570\u586b\u5145\u7ec4\u5408\u8868\u73b0\u6700\u4f18，\u6d4b\u8bd5\u51c6\u786e\u7387\u8fbe99%，\u4e14\u5404\u9879\u6307\u6807（\u7cbe\u786e\u7387\u3001\u53ec\u56de\u7387\u3001F1\u5206\u6570）\u5747\u63a5\u8fd1\u6ee1\u5206\u3002<\/p>\n
\u6700\u4f18\u7ec4\u5408\u5408\u7406\u6027\u5206\u6790<\/h5>\n
- \n
\u4e2d\u4f4d\u6570\u586b\u5145\u7684\u4f18\u52bf：\u77ff\u7269\u6210\u5206\u6570\u636e\u53ef\u80fd\u5b58\u5728\u6781\u7aef\u503c（\u5982\u6742\u8d28\u542b\u91cf\u6ce2\u52a8），\u4e2d\u4f4d\u6570\u76f8\u6bd4\u5747\u503c\u66f4\u6297\u5e72\u6270，\u80fd\u4fdd\u7559\u6570\u636e\u7684\u771f\u5b9e\u5206\u5e03，\u4e3a\u6a21\u578b\u63d0\u4f9b\u7a33\u5b9a\u8f93\u5165；\u540c\u65f6\u4e0d\u5220\u9664\u6837\u672c，\u4fdd\u7559\u4e86\u66f4\u591a\u8bad\u7ec3\u6570\u636e，\u5c24\u5176\u9002\u5408\u5c0f\u6837\u672c\u573a\u666f\u3002<\/p>\n<\/li>\n
- \n
XGBoost\u7684\u4f18\u52bf：\u77ff\u7269\u6210\u5206\u7279\u5f81\u95f4\u5b58\u5728\u590d\u6742\u7684\u975e\u7ebf\u6027\u5173\u8054，XGBoost\u80fd\u6709\u6548\u6355\u6349\u8fd9\u4e9b\u5173\u7cfb，\u4e14\u901a\u8fc7\u6b63\u5219\u5316（gamma\u3001subsample）\u907f\u514d\u8fc7\u62df\u5408，\u5728\u591a\u5206\u7c7b\u4efb\u52a1\u4e2d\u7cbe\u5ea6\u4f18\u4e8e\u4f20\u7edf\u6a21\u578b；\u4f18\u5316\u540e\u7684\u6807\u7b7e\u5904\u7406\u903b\u8f91\u8fdb\u4e00\u6b65\u786e\u4fdd\u4e86\u6a21\u578b\u7a33\u5b9a\u6027\u3002<\/p>\n<\/li>\n
\u5176\u4ed6\u65b9\u6848\u8868\u73b0\u5bf9\u6bd4<\/h5>\n
- \n
- \n
\u586b\u5145\u65b9\u6cd5\u5c42\u9762：\u57fa\u4e8e\u6a21\u578b\u7684\u586b\u5145（\u7ebf\u6027\u56de\u5f52\u3001\u968f\u673a\u68ee\u6797）\u6574\u4f53\u4f18\u4e8e\u7b80\u5355\u586b\u5145，\u4f46\u8ba1\u7b97\u6210\u672c\u66f4\u9ad8；\u5220\u9664\u7a7a\u6570\u636e\u884c\u56e0\u4e22\u5931\u6837\u672c，\u7cbe\u5ea6\u666e\u904d\u504f\u4f4e\u3002<\/p>\n<\/li>\n
- \n
\u6a21\u578b\u5c42\u9762：\u96c6\u6210\u5b66\u4e60（XGBoost\u3001\u968f\u673a\u68ee\u6797）\u6574\u4f53\u4f18\u4e8e\u4f20\u7edf\u673a\u5668\u5b66\u4e60\u4e0e\u6df1\u5ea6\u5b66\u4e60；MLP\u3001CNN\u8868\u73b0\u7565\u900a，\u53ef\u80fd\u56e0\u6570\u636e\u96c6\u7279\u5f81\u7ef4\u5ea6\u4e0d\u9ad8，\u6df1\u5ea6\u5b66\u4e60\u7684\u4f18\u52bf\u96be\u4ee5\u53d1\u6325\u3002<\/p>\n<\/li>\n<\/ul>\n
\u4e09\u3001\u9879\u76ee\u5173\u952e\u95ee\u9898\u4e0e\u89e3\u51b3\u65b9\u6848<\/h3>\n
\u590d\u73b0\u8fc7\u7a0b\u4e2d\u9047\u5230\u591a\u4e2a\u5178\u578b\u95ee\u9898，\u9488\u5bf9\u6027\u89e3\u51b3\u65b9\u6cd5\u53ef\u4e3a\u540c\u7c7b\u9879\u76ee\u63d0\u4f9b\u53c2\u8003：<\/p>\n
- \n
SVM\u8bad\u7ec3\u5361\u4f4f：Poly\u6838+\u9ad8degree+probability=True\u5bfc\u81f4\u8ba1\u7b97\u91cf\u66b4\u589e，\u4f18\u5316\u53c2\u6570（\u964d\u4f4edegree\u3001\u5173\u95edprobability）+\u589e\u52a0\u7f13\u5b58，\u89e3\u51b3\u5361\u987f\u95ee\u9898\u3002<\/p>\n<\/li>\n
- \n
XGBoost\u6807\u7b7e\u62a5\u9519：\u6807\u7b7e\u975e\u8fde\u7eed\u6574\u6570\u5bfc\u81f4\u591a\u5206\u7c7b\u5931\u8d25，\u6dfb\u52a0\u52a8\u6001\u6807\u7b7e\u6620\u5c04，\u5c06\u539f\u59cb\u6807\u7b7e\u8f6c\u4e3a\u4ece0\u5f00\u59cb\u7684\u8fde\u7eed\u6574\u6570，\u540c\u65f6\u52a8\u6001\u8bbe\u7f6e\u7c7b\u522b\u6570\u3002<\/p>\n<\/li>\n
- \n
\u6a21\u578b\u8bad\u7ec3\u5185\u5b58\u8fc7\u8f7d：\u903b\u8f91\u56de\u5f52\u3001XGBoost\u542f\u7528\u591a\u7ebf\u7a0b\u65f6\u5185\u5b58\u4e0d\u8db3，\u7981\u7528\u591a\u7ebf\u7a0b\u6216\u9650\u5236\u6838\u5fc3\u6570，\u4f18\u5148\u4fdd\u8bc1\u7a33\u5b9a\u6027\u3002<\/p>\n<\/li>\n
- \n
\u6df1\u5ea6\u5b66\u4e60\u8bbe\u5907\u9002\u914d：\u81ea\u52a8\u68c0\u6d4bGPU\/CPU，\u5c06\u5f20\u91cf\u8f6c\u79fb\u81f3\u5bf9\u5e94\u8bbe\u5907，\u8bad\u7ec3\u65f6\u7528torch.no_grad()\u5173\u95ed\u68af\u5ea6\u8ba1\u7b97，\u51cf\u5c11\u5185\u5b58\u5360\u7528\u3002<\/p>\n<\/li>\n
\u56db\u3001\u603b\u7ed3\u4e0e\u5c55\u671b<\/h3>\n
4.1 \u9879\u76ee\u6536\u83b7<\/h4>\n
\u672c\u9879\u76ee\u901a\u8fc7\u7cfb\u7edf\u5bf9\u6bd46\u79cd\u586b\u5145\u65b9\u6cd5\u4e0e7\u79cd\u5206\u7c7b\u6a21\u578b，\u9a8c\u8bc1\u4e86\u201c\u4e2d\u4f4d\u6570\u586b\u5145+XGBoost\u201d\u5728\u77ff\u7269\u5206\u7c7b\u4efb\u52a1\u4e2d\u7684\u4f18\u8d8a\u6027，\u540c\u65f6\u68b3\u7406\u4e86\u6570\u636e\u9884\u5904\u7406\u3001\u6a21\u578b\u4f18\u5316\u3001\u95ee\u9898\u6392\u67e5\u7684\u5b8c\u6574\u6d41\u7a0b\u3002\u6838\u5fc3\u7ed3\u8bba：\u6570\u636e\u9884\u5904\u7406\u7684\u8d28\u91cf\u76f4\u63a5\u51b3\u5b9a\u6a21\u578b\u4e0a\u9650，\u9488\u5bf9\u6570\u636e\u7279\u70b9\u9009\u62e9\u586b\u5145\u65b9\u6cd5，\u6bd4\u76f2\u76ee\u8ffd\u6c42\u590d\u6742\u6a21\u578b\u66f4\u91cd\u8981\u3002<\/p>\n
4.2 \u540e\u7eed\u4f18\u5316\u65b9\u5411<\/h4>\n
- \n
- \n
\u6a21\u578b\u90e8\u7f72：\u5c06\u6700\u4f18XGBoost\u6a21\u578b\u5c01\u88c5\u4e3a\u9884\u6d4b\u63a5\u53e3，\u652f\u6301\u65b0\u77ff\u7269\u6837\u672c\u7684\u5b9e\u65f6\u5206\u7c7b；<\/p>\n<\/li>\n
- \n
\u7279\u5f81\u5de5\u7a0b：\u901a\u8fc7XGBoost\u7684\u7279\u5f81\u91cd\u8981\u6027\u5206\u6790，\u7b5b\u9009\u5173\u952e\u77ff\u7269\u6210\u5206\u7279\u5f81，\u7b80\u5316\u6a21\u578b\u7ed3\u6784；<\/p>\n<\/li>\n
- \n
\u6a21\u578b\u878d\u5408：\u7ed3\u5408XGBoost\u4e0e\u968f\u673a\u68ee\u6797\u7684\u9884\u6d4b\u7ed3\u679c，\u8fdb\u4e00\u6b65\u63d0\u5347\u5206\u7c7b\u7cbe\u5ea6；<\/p>\n<\/li>\n
- \n
\u6570\u636e\u6269\u5145：\u9488\u5bf9\u5c0f\u6837\u672c\u573a\u666f，\u901a\u8fc7\u6570\u636e\u589e\u5f3a\u6280\u672f\u6269\u5145\u8bad\u7ec3\u96c6，\u63d0\u5347\u6a21\u578b\u6cdb\u5316\u80fd\u529b\u3002<\/p>\n<\/li>\n<\/ul>\n
\u77ff\u7269\u5206\u7c7b\u7cfb\u7edf\u7684\u590d\u73b0，\u4e0d\u4ec5\u5b9e\u73b0\u4e86\u9ad8\u7cbe\u5ea6\u5206\u7c7b\u76ee\u6807，\u66f4\u9a8c\u8bc1\u4e86\u673a\u5668\u5b66\u4e60\u6280\u672f\u5728\u5730\u8d28\u9886\u57df\u7684\u5b9e\u7528\u6027\u3002\u540e\u7eed\u53ef\u7ed3\u5408\u66f4\u591a\u5b9e\u9645\u573a\u666f\u9700\u6c42，\u6301\u7eed\u4f18\u5316\u6a21\u578b\u4e0e\u6d41\u7a0b，\u4e3a\u77ff\u4ea7\u52d8\u63a2\u3001\u5730\u8d28\u5206\u6790\u63d0\u4f9b\u66f4\u9ad8\u6548\u7684\u6280\u672f\u652f\u6491\u3002<\/p>\n
\u4ee3\u7801\u90e8\u5206\u89e3\u6790：<\/h4>\n
\u8868\u683c\u5408\u5e76：<\/h5>\n
\u6bd4\u8f83\u53cd\u76f4\u89c9，\u884c\u5408\u5e76\u6307\u7684\u662f\u628a\u4e24\u4e2a\u8868\u683c\u4e00\u4e0a\u4e00\u4e0b\u62fc\u63a5\u5728\u4e00\u8d77，axis=0；\u5217\u5408\u5e76\u6307\u7684\u662f\u4e00\u5de6\u4e00\u53f3\u822a\u5411\u62fc\u63a5,axis=1\u3002<\/p>\n
![\"\"](\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2026\/01\/20260116165710-696a6de67b6f5.png\")
<\/h4>\n<\/h4>\n
\u586b\u5145\u7a7a\u7f3a\u503c：
![\"\"](\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2026\/01\/20260116165710-696a6de68eba7.png\")
<\/h5>\n","protected":false},"excerpt":{"rendered":"\u5728\u5730\u8d28\u52d8\u63a2\u3001\u77ff\u4ea7\u5f00\u53d1\u7b49\u9886\u57df，\u77ff\u7269\u5206\u7c7b\u7684\u51c6\u786e\u6027\u76f4\u63a5\u5f71\u54cd\u540e\u7eed\u5206\u6790\u51b3\u7b56\u3002\u4f20\u7edf\u4eba\u5de5\u5206\u7c7b\u4f9d\u8d56\u4e13\u4e1a\u7ecf\u9a8c，\u6548\u7387\u4f4e\u4e14\u4e3b\u89c2\u6027\u5f3a，\u800c\u673a\u5668\u5b66\u4e60\u4e0e\u6df1\u5ea6\u5b66\u4e60\u6280\u672f\u80fd\u5b9e\u73b0\u81ea\u52a8\u5316\u3001\u9ad8\u7cbe\u5ea6\u5206\u7c7b\u3002\u672c\u6587\u5c06\u8be6\u7ec6\u8bb0\u5f55\u590d\u73b0\u77ff\u7269\u5206\u7c7b\u7cfb\u7edf\u7684\u5168\u8fc7\u7a0b，\u6db5\u76d66\u79cd\u6570\u636e\u586b\u5145\u65b9\u6cd5\u30017\u79cd\u5206\u7c7b\u6a21\u578b\u7684\u642d\u914d\u5b9e\u9a8c，\u6700\u7ec8\u7b5b\u9009\u51fa\u6700\u4f18\u65b9\u6848，\u51c6\u786e\u7387\u8fbe99%\u3002\u4e00\u3001\u9879\u76ee\u80cc\u666f\u4e0e\u6280\u672f\u68081.1 \u9879\u76ee\u76ee\u6807\u57fa\u4e8e\u77ff\u7269\u6210\u5206\u7279\u5f81\u6570\u636e，\u6784\u5efa\u5206\u7c7b\u7cfb\u7edf，\u5b9e\u73b0\u5bf9\u4e0d\u540c\u7c7b\u578b<\/p>\n","protected":false},"author":2,"featured_media":61107,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[50,2395,62],"topic":[],"class_list":["post-61110","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-server","tag-50","tag-2395","tag-62"],"yoast_head":"\n
\u77ff\u7269\u5206\u7c7b\u7cfb\u7edf\u8bbe\u8ba1 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.wsisp.com\/helps\/61110.html\" \/>\n<meta property=\"og:locale\" content=\"zh_CN\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"\u77ff\u7269\u5206\u7c7b\u7cfb\u7edf\u8bbe\u8ba1 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"og:description\" content=\"\u5728\u5730\u8d28\u52d8\u63a2\u3001\u77ff\u4ea7\u5f00\u53d1\u7b49\u9886\u57df，\u77ff\u7269\u5206\u7c7b\u7684\u51c6\u786e\u6027\u76f4\u63a5\u5f71\u54cd\u540e\u7eed\u5206\u6790\u51b3\u7b56\u3002\u4f20\u7edf\u4eba\u5de5\u5206\u7c7b\u4f9d\u8d56\u4e13\u4e1a\u7ecf\u9a8c，\u6548\u7387\u4f4e\u4e14\u4e3b\u89c2\u6027\u5f3a，\u800c\u673a\u5668\u5b66\u4e60\u4e0e\u6df1\u5ea6\u5b66\u4e60\u6280\u672f\u80fd\u5b9e\u73b0\u81ea\u52a8\u5316\u3001\u9ad8\u7cbe\u5ea6\u5206\u7c7b\u3002\u672c\u6587\u5c06\u8be6\u7ec6\u8bb0\u5f55\u590d\u73b0\u77ff\u7269\u5206\u7c7b\u7cfb\u7edf\u7684\u5168\u8fc7\u7a0b，\u6db5\u76d66\u79cd\u6570\u636e\u586b\u5145\u65b9\u6cd5\u30017\u79cd\u5206\u7c7b\u6a21\u578b\u7684\u642d\u914d\u5b9e\u9a8c，\u6700\u7ec8\u7b5b\u9009\u51fa\u6700\u4f18\u65b9\u6848，\u51c6\u786e\u7387\u8fbe99%\u3002\u4e00\u3001\u9879\u76ee\u80cc\u666f\u4e0e\u6280\u672f\u68081.1 \u9879\u76ee\u76ee\u6807\u57fa\u4e8e\u77ff\u7269\u6210\u5206\u7279\u5f81\u6570\u636e，\u6784\u5efa\u5206\u7c7b\u7cfb\u7edf，\u5b9e\u73b0\u5bf9\u4e0d\u540c\u7c7b\u578b\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.wsisp.com\/helps\/61110.html\" \/>\n<meta property=\"og:site_name\" content=\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"article:published_time\" content=\"2026-01-16T16:57:11+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2026\/01\/20260116165710-696a6de656a8e.png\" \/>\n<meta name=\"author\" content=\"admin\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u4f5c\u8005\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin\" \/>\n\t<meta name=\"twitter:label2\" content=\"\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4\" \/>\n\t<meta name=\"twitter:data2\" content=\"27 \u5206\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/61110.html\",\"url\":\"https:\/\/www.wsisp.com\/helps\/61110.html\",\"name\":\"\u77ff\u7269\u5206\u7c7b\u7cfb\u7edf\u8bbe\u8ba1 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"isPartOf\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\"},\"datePublished\":\"2026-01-16T16:57:11+00:00\",\"dateModified\":\"2026-01-16T16:57:11+00:00\",\"author\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\"},\"breadcrumb\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/61110.html#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.wsisp.com\/helps\/61110.html\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/61110.html#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"\u9996\u9875\",\"item\":\"https:\/\/www.wsisp.com\/helps\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"\u77ff\u7269\u5206\u7c7b\u7cfb\u7edf\u8bbe\u8ba1\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\",\"url\":\"https:\/\/www.wsisp.com\/helps\/\",\"name\":\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"description\":\"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\",\"name\":\"admin\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"contentUrl\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"caption\":\"admin\"},\"sameAs\":[\"http:\/\/wp.wsisp.com\"],\"url\":\"https:\/\/www.wsisp.com\/helps\/author\/admin\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"\u77ff\u7269\u5206\u7c7b\u7cfb\u7edf\u8bbe\u8ba1 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.wsisp.com\/helps\/61110.html","og_locale":"zh_CN","og_type":"article","og_title":"\u77ff\u7269\u5206\u7c7b\u7cfb\u7edf\u8bbe\u8ba1 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","og_description":"\u5728\u5730\u8d28\u52d8\u63a2\u3001\u77ff\u4ea7\u5f00\u53d1\u7b49\u9886\u57df，\u77ff\u7269\u5206\u7c7b\u7684\u51c6\u786e\u6027\u76f4\u63a5\u5f71\u54cd\u540e\u7eed\u5206\u6790\u51b3\u7b56\u3002\u4f20\u7edf\u4eba\u5de5\u5206\u7c7b\u4f9d\u8d56\u4e13\u4e1a\u7ecf\u9a8c，\u6548\u7387\u4f4e\u4e14\u4e3b\u89c2\u6027\u5f3a，\u800c\u673a\u5668\u5b66\u4e60\u4e0e\u6df1\u5ea6\u5b66\u4e60\u6280\u672f\u80fd\u5b9e\u73b0\u81ea\u52a8\u5316\u3001\u9ad8\u7cbe\u5ea6\u5206\u7c7b\u3002\u672c\u6587\u5c06\u8be6\u7ec6\u8bb0\u5f55\u590d\u73b0\u77ff\u7269\u5206\u7c7b\u7cfb\u7edf\u7684\u5168\u8fc7\u7a0b，\u6db5\u76d66\u79cd\u6570\u636e\u586b\u5145\u65b9\u6cd5\u30017\u79cd\u5206\u7c7b\u6a21\u578b\u7684\u642d\u914d\u5b9e\u9a8c，\u6700\u7ec8\u7b5b\u9009\u51fa\u6700\u4f18\u65b9\u6848，\u51c6\u786e\u7387\u8fbe99%\u3002\u4e00\u3001\u9879\u76ee\u80cc\u666f\u4e0e\u6280\u672f\u68081.1 \u9879\u76ee\u76ee\u6807\u57fa\u4e8e\u77ff\u7269\u6210\u5206\u7279\u5f81\u6570\u636e，\u6784\u5efa\u5206\u7c7b\u7cfb\u7edf，\u5b9e\u73b0\u5bf9\u4e0d\u540c\u7c7b\u578b","og_url":"https:\/\/www.wsisp.com\/helps\/61110.html","og_site_name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","article_published_time":"2026-01-16T16:57:11+00:00","og_image":[{"url":"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2026\/01\/20260116165710-696a6de656a8e.png"}],"author":"admin","twitter_card":"summary_large_image","twitter_misc":{"\u4f5c\u8005":"admin","\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4":"27 \u5206"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.wsisp.com\/helps\/61110.html","url":"https:\/\/www.wsisp.com\/helps\/61110.html","name":"\u77ff\u7269\u5206\u7c7b\u7cfb\u7edf\u8bbe\u8ba1 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","isPartOf":{"@id":"https:\/\/www.wsisp.com\/helps\/#website"},"datePublished":"2026-01-16T16:57:11+00:00","dateModified":"2026-01-16T16:57:11+00:00","author":{"@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41"},"breadcrumb":{"@id":"https:\/\/www.wsisp.com\/helps\/61110.html#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.wsisp.com\/helps\/61110.html"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/www.wsisp.com\/helps\/61110.html#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"\u9996\u9875","item":"https:\/\/www.wsisp.com\/helps"},{"@type":"ListItem","position":2,"name":"\u77ff\u7269\u5206\u7c7b\u7cfb\u7edf\u8bbe\u8ba1"}]},{"@type":"WebSite","@id":"https:\/\/www.wsisp.com\/helps\/#website","url":"https:\/\/www.wsisp.com\/helps\/","name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","description":"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"zh-Hans"},{"@type":"Person","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41","name":"admin","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/","url":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","contentUrl":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","caption":"admin"},"sameAs":["http:\/\/wp.wsisp.com"],"url":"https:\/\/www.wsisp.com\/helps\/author\/admin"}]}},"_links":{"self":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/61110","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/comments?post=61110"}],"version-history":[{"count":0,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/61110\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media\/61107"}],"wp:attachment":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media?parent=61110"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/categories?post=61110"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/tags?post=61110"},{"taxonomy":"topic","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/topic?post=61110"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
- \n
- \n
- \n
- \n
- \n