糖尿病康复 > 讯飞——糖尿病遗传风险检测挑战赛解决方案

讯飞——糖尿病遗传风险检测挑战赛解决方案

时间：2022-09-16 14:31:33

0. 赛事背景1. 读取数据2. 数据探索及预处理2.1 缺失值2.2 分析字段类型2.3 计算字段相关性3. 特征工程3.1 特征构造4. 模型训练4.1 LightGBM (0.96206)4.2 随机森林（0.96324）4.3 XGBoost (0.95981)4.4 CatBoost（0.95854）4.5 AdaBoost（0.96098）4.6 集成模型（0.95971）4.7 Stacking（0.96577）4.8 归一化数据，pytorch神经网络4.9 SVM4.10 sklearn神经网络5. 总结思考

0. 赛事背景

截至，中国糖尿病患者近1.3亿。中国糖尿病患病原因受生活方式、老龄化、城市化、家族遗传等多种因素影响。同时，糖尿病患者趋向年轻化。

糖尿病可导致心血管、肾脏、脑血管并发症的发生。因此，准确诊断出患有糖尿病个体具有非常重要的临床意义。糖尿病早期遗传风险预测将有助于预防糖尿病的发生。

根据《中国2型糖尿病防治指南（版）》，糖尿病的诊断标准是具有典型糖尿病症状（烦渴多饮、多尿、多食、不明原因的体重下降）且随机静脉血浆葡萄糖≥11.1mmol/L或空腹静脉血浆葡萄糖≥7.0mmol/L或口服葡萄糖耐量试验（OGTT）负荷后2h血浆葡萄糖≥11.1mmol/L。

在这次比赛中，您需要通过训练数据集构建糖尿病遗传风险预测模型，然后预测出测试数据集中个体是否患有糖尿病，和我们一起帮助糖尿病患者解决这“甜蜜的烦恼”。

训练集说明

训练集（比赛训练集.csv）一共有5070条数据，用于构建您的预测模型（您可能需要先进行数据分析）。数据的字段有编号、性别、出生年份、体重指数、糖尿病家族史、舒张压、口服耐糖量测试、胰岛素释放实验、肱三头肌皮褶厚度、患有糖尿病标识（最后一列），您也可以通过特征工程技术构建新的特征。

测试集说明

测试集（比赛测试集.csv）一共有1000条数据，用于验证预测模型的性能。数据的字段有编号、性别、出生年份、体重指数、糖尿病家族史、舒张压、口服耐糖量测试、胰岛素释放实验、肱三头肌皮褶厚度。

评估指标

对于提交的结果，系统会采用二分类任务中的F1-score指标进行评价，F1-score越大说明预测模型性能越好。

本质为二分类预测，不是很难，注意做好特征工程，提升数据质量。

比赛详情见官网。

1. 读取数据

import pandas as pdimport numpy as npimport matplotlib.pyplot as pltimport lightgbmtrain_data = pd.read_csv('比赛训练集.csv',encoding='gbk')test_data = pd.read_csv('比赛测试集.csv',encoding='gbk')train_data.describe()

患有糖尿病标识的均值为0.38，说明未患病：患病 ≈6:4\approx 6 : 4≈6:4，数据还算较为平衡。

test_data.describe()

2. 数据探索及预处理

2.1 缺失值

统计每个字段的缺失比例，并进行填充。可以看到舒张压指标的缺失值较多，用字段均值将其填充。

print('训练集各字段缺失比例：')print(train_data.isnull().mean(0))print('\n测试集各字段缺失比例：')print(test_data.isnull().mean(0))# 用均值填充缺失值train_data['舒张压'] = train_data['舒张压'].fillna(train_data['舒张压'].mean())test_data['舒张压'] = test_data['舒张压'].fillna(test_data['舒张压'].mean())

训练集各字段缺失比例：编号0.000000性别0.000000出生年份 0.000000体重指数 0.000000糖尿病家族史0.000000舒张压 0.048718口服耐糖量测试0.000000胰岛素释放实验0.000000肱三头肌皮褶厚度 0.000000患有糖尿病标识0.000000dtype: float64测试集各字段缺失比例：编号0.000性别0.000出生年份 0.000体重指数 0.000糖尿病家族史0.000舒张压 0.049口服耐糖量测试0.000胰岛素释放实验0.000肱三头肌皮褶厚度 0.000dtype: float64

2.2 分析字段类型

print(train_data.columns)train_data.describe()

Index(['编号', '性别', '出生年份', '体重指数', '糖尿病家族史', '舒张压', '口服耐糖量测试', '胰岛素释放实验','肱三头肌皮褶厚度', '患有糖尿病标识'],dtype='object')

编号与是否患病没关系，删除;

性别为类别变量，只有0，1，不再需要进行编码；

糖尿病家族病史为文本型变量，需要转化为数值变量；

其他均为数值型变量，可以暂时不变。

train_data = train_data.drop(['编号'], axis=1)test_data = test_data.drop(['编号'], axis=1)

2.3 计算字段相关性

查看各字段之间的相关性，防止多重共线性。

Ref:
Python绘制相关性热力图

train_corr = train_data.drop('糖尿病家族史',axis=1).corr()import seaborn as snsplt.subplots(figsize=(9,9),dpi=80,facecolor='w') # 设置画布大小，分辨率，和底色plt.rcParams['font.sans-serif'] = ['SimHei'] # 黑体plt.rcParams['axes.unicode_minus'] = False # 解决无法显示符号的问题sns.set(font='SimHei', font_scale=0.8) # 解决Seaborn中文显示问题#annot为热力图上显示数据；fmt='.2g'为数据保留两位有效数字,square呈现正方形，vmax最大值为1fig=sns.heatmap(train_corr,annot=True, vmax=1, square=True, cmap="Blues", fmt='.2g')#保存图片fig.get_figure().savefig('train_corr.png',bbox_inches='tight',transparent=True)#bbox_inches让图片显示完整，transparent=True让图片背景透明

这里中文显示出了点问题，但是可以看到各特征变量与是否患病没有显著线性关系（但可能有非线性关系），各特征变量之间也不存在多重共线性，可以继续下一步操作。

3. 特征工程

这一步至关重要，主要是有两个目的：

特征构造：尝试构建有价值的新变量；特征筛选：删除对因变量影响不大的冗余变量。

由于这里的特征也不是很多，就不做筛选了。

3.1 特征构造

可以用统计指标，已有知识、经验构造新的变量，具体到这个问题上可以有BMI指数、舒张压范围、年龄等。

特征构造方法：

特征的统计指标；特征之间的四则运算；交叉特征；分解类别特征。如将三个颜色分解为“知道颜色”和“不知道颜色”。特征分箱。将数值型特征变量按段划分，得到类别型特征。重构特征。单位转换、整数部分与小数部分分离等。根据已有经验构造新的特征变量，比如xx因子。

Ref:
[1] 深度了解特征工程

# 将出生年份换算成年龄train_data['年龄']=-train_data['出生年份'] #换成年龄test_data['年龄']=-test_data['出生年份']train_data = train_data.drop('出生年份', axis=1)test_data = test_data.drop('出生年份', axis=1)

# 家族史转换，方法一，label编码from sklearn.preprocessing import OneHotEncoder, LabelEncoderdef FHOD(a):if a=='无记录':return 0elif a=='叔叔或者姑姑有一方患有糖尿病' or a=='叔叔或姑姑有一方患有糖尿病':return 1else:return 2train_data['糖尿病家族史'] = train_data['糖尿病家族史'].apply(FHOD)test_data['糖尿病家族史'] = test_data['糖尿病家族史'].apply(FHOD)# history = train_data['糖尿病家族史']# print(set(history))# history.loc[history=='叔叔或姑姑有一方患有糖尿病'] = '叔叔或者姑姑有一方患有糖尿病'# le = LabelEncoder()# h = le.fit_transform(history)# 方法二，onehot 编码# def onehot_transform(data):## 将家族史的文本型变量转换为onehot编码。#onehot = OneHotEncoder()#data.loc[data['糖尿病家族史']=='叔叔或姑姑有一方患有糖尿病', '糖尿病家族史'] = '叔叔或者姑姑有一方患有糖尿病'#data_onehot = pd.DataFrame(onehot.fit_transform(data[['糖尿病家族史']]).toarray(),#columns=onehot.get_feature_names(['糖尿病家族史']), dtype='int32')#return data_onehot# data_train_history = onehot_transform(train_data)# data_test_history = onehot_transform(test_data)

def BMI(a):"""人体的成人体重指数正常值是在18.5-24之间低于18.5是体重指数过轻在24-27之间是体重超重27以上考虑是肥胖高于32了就是非常的肥胖。"""if a<18.5:return 0elif 18.5<=a<=24:return 1elif 24<a<=27:return 2elif 27<a<=32:return 3else:return 4train_data['BMI']=train_data['体重指数'].apply(BMI)test_data['BMI']=test_data['体重指数'].apply(BMI)# 转换舒张压为类别型变量def DBP(a):# 舒张压范围为60-90if a<60:return 0elif 60<=a<=90:return 1elif a>90:return 2else:return atrain_data['DBP'] = train_data['舒张压'].apply(DBP)test_data['DBP'] = test_data['舒张压'].apply(DBP)X_train = train_data.drop('患有糖尿病标识', axis=1)Y_train = train_data['患有糖尿病标识']X_train['年龄'] = X_train['年龄'].astype(float)X_test = test_datatrain_data.describe()

4. 模型训练

可用于分类问题的模型非常丰富，常见的如下图：

4.1 LightGBM (0.96206)

首先尝试构建LightGBM模型。

Ref:
[1] Lightgbm原理、参数详解及python实例
[2] 深入理解LightGBM

以下为lightgbm采用5折交叉训练的代码：

#使用Lightgbm方法训练数据集，使用5折交叉验证的方法获得5个测试集预测结果from sklearn.model_selection import KFoldfrom sklearn.model_selection import StratifiedKFold, GridSearchCVdef select_by_lgb(train_data,train_label,test_data,random_state=1234, n_splits=5,metric='auc',num_round=10000,early_stopping_rounds=100):# kfold = KFold(n_splits=n_splits, shuffle=True, random_state=random_state)kfold = StratifiedKFold(n_splits=n_splits, shuffle=True, random_state=random_state)fold=0result0 = []for train_idx, val_idx in kfold.split(train_data, train_label):random_state+=1train_x = train_data.loc[train_idx]train_y = train_label.loc[train_idx]test_x = train_data.loc[val_idx]test_y = train_label.loc[val_idx]clf = lightgbmtrain_matrix=clf.Dataset(train_x,label=train_y)test_matrix=clf.Dataset(test_x,label=test_y)params={'boosting_type': 'gbdt','objective': 'binary','learning_rate': 0.1,# 'max_depth': 7,# 'num_leaves': 10,'metric': metric,'seed': random_state,'silent': True,'nthread':-1 }model=clf.train(params,train_matrix,num_round,valid_sets=test_matrix,early_stopping_rounds=early_stopping_rounds)pre_y=model.predict(test_data)result0.append(pre_y)fold+=1pred_test = pd.DataFrame(result0).T# 将5次预测结果求平均值pred_test['average'] = pred_test.mean(axis=1)#因为竞赛需要你提交最后的预测判断，而模型给出的预测结果是概率，因此我们认为概率>0.5的即该患者有糖尿病，概率<=0.5的没有糖尿病pred_test['label'] = pred_test['average'].apply(lambda x:1 if x>0.5 else 0)## 导出结果result = pd.read_csv('提交示例.csv')result['label']=pred_test['label']return result

后面其他模型也需要进行k折交叉训练，这里定义一个k折交叉训练的函数，方便后续调用。

from sklearn.model_selection import KFold, StratifiedKFoldfrom sklearn.metrics import roc_auc_score, f1_scoredef SKFold(train_data,train_label,test_data, model, random_state=1234, n_splits=5,metric='auc',num_round=10000,early_stopping_rounds=100):# 采用分层K折交叉验证训练模型。kfold = StratifiedKFold(n_splits=n_splits, shuffle=True, random_state=random_state)fold = 1pred_test = []for train_idx, val_idx in kfold.split(train_data, train_label):random_state+=1train_x = train_data.loc[train_idx]train_y = train_label.loc[train_idx]val_x = train_data.loc[val_idx]val_y = train_label.loc[val_idx]eval_set = (val_x, val_y)clf = modelmodel_trained = clf.fit(train_x, train_y)# model_trained = clf.fit(train_x,train_y,early_stopping_rounds=early_stopping_rounds, verbose=False)# model_trained = clf.fit(train_x, train_y, eval_set=eval_set, early_stopping_rounds=early_stopping_rounds)pre_y = model_trained.predict(test_data)pred_test.append(pre_y)auc_train = roc_auc_score(train_y, model_trained.predict(train_x))auc_val = roc_auc_score(val_y, model_trained.predict(val_x))f_score_train = f1_score(train_y, model_trained.predict(train_x))f_score_val = f1_score(val_y, model_trained.predict(val_x))print('Fold: %d, AUC_train: %.4f, AUC_val: %.4f, F1-score_train: %.4f, F1-score_val: %.4f'%(fold, auc_train, auc_val, f_score_train, f_score_val))fold += 1pred_test = pd.DataFrame(pred_test).T# 将5次预测结果求平均值pred_test['average'] = pred_test.mean(axis=1)#因为竞赛需要你提交最后的预测判断，而模型给出的预测结果是概率，因此我们认为概率>0.5的即该患者有糖尿病，概率<=0.5的没有糖尿病pred_test['label'] = pred_test['average'].apply(lambda x:1 if x>0.5 else 0)## 导出结果result=pd.read_csv('提交示例.csv')result['label']=pred_test['label']return result

由于比赛的测试集未公布，我们只能提交预测结果然后得到测试集上的分数，这里以表现较好的lightgbm模型作为baseline，若与lightgbm的预测结果相差较多则说明该模型表现不行。

def evaluate(result_LightGBM, result_others):# 以lightGBM的结果为基准，评估其他模型的表现。c = result_LightGBM['label'] - result_others['label']count = 0for i in c:if i != 0:count += 1print('与LightGBM预测不同的样本数： ', count)print(c[c!=0])return count

先用select_by_lgb快速跑出一个baseline，然后用网格搜索得到最优参数，接着用最优参数组合在训练一遍模型，最后将结果提交。

random_state = 1234result_LightGBM = select_by_lgb(X_train, Y_train, X_test) #baselineresult_LightGBM.to_csv('result_lightGBM.csv',index=False)# 试试网格搜索最优参数import lightgbm as lgbparams_test = {'max_depth': range(4, 10, 1),'num_leaves': range(10, 60, 10)}skf = StratifiedKFold(n_splits=5) gsearch1 = GridSearchCV(estimator=lgb.LGBMClassifier(boosting_type='gbdt',objective='binary',metrics='auc',learning_rate=0.1, n_estimators=325, max_depth=8, bagging_fraction = 0.8,feature_fraction = 0.8), param_grid=params_test,scoring='roc_auc', cv=skf, n_jobs=-1)gsearch1.fit(X_train, Y_train)print(gsearch1.best_params_)print(gsearch1.best_score_)# 用最优参数再训练一遍model_lgb = lgb.LGBMClassifier(boosting_type='gbdt',objective='binary',metrics='auc',learning_rate=0.1, n_estimators=200, num_leaves=10, silent=True,max_depth=7)result_SKFold_lgb = SKFold(X_train, Y_train, X_test, model_lgb, n_splits=5)result_SKFold_lgb.to_csv('result_SKFold_lgb.csv',index=False)diff_lgb = evaluate(result_LightGBM, result_SKFold_lgb)

[LightGBM] [Warning] Unknown parameter: silent[LightGBM] [Warning] Unknown parameter: silent[LightGBM] [Info] Number of positive: 1549, number of negative: 2507[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000386 seconds.You can set `force_row_wise=true` to remove the overhead.And if memory is not enough, you can set `force_col_wise=true`.[LightGBM] [Info] Total Bins 1047[LightGBM] [Info] Number of data points in the train set: 4056, number of used features: 10[LightGBM] [Warning] Unknown parameter: silent[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.381903 -> initscore=-0.481477[LightGBM] [Info] Start training from score -0.481477[1]valid_0's auc: 0.979322Training until validation scores don't improve for 100 rounds[2]valid_0's auc: 0.980354[3]valid_0's auc: 0.982351[4]valid_0's auc: 0.981993Early stopping, best iteration is:[40]valid_0's auc: 0.989851[LightGBM] [Warning] feature_fraction is set=0.8, colsample_bytree=1.0 will be ignored. Current value: feature_fraction=0.8[LightGBM] [Warning] bagging_fraction is set=0.8, subsample=1.0 will be ignored. Current value: bagging_fraction=0.8{'max_depth': 7, 'num_leaves': 10}0.9902553653233458Fold: 1, AUC_train: 0.9917, AUC_val: 0.9461, F1-score_train: 0.9909, F1-score_val: 0.9358[LightGBM] [Warning] feature_fraction is set=0.8, colsample_bytree=1.0 will be ignored. Current value: feature_fraction=0.8[LightGBM] [Warning] bagging_fraction is set=0.8, subsample=1.0 will be ignored. Current value: bagging_fraction=0.8Fold: 2, AUC_train: 0.9852, AUC_val: 0.9487, F1-score_train: 0.9837, F1-score_val: 0.9386[LightGBM] [Warning] feature_fraction is set=0.8, colsample_bytree=1.0 will be ignored. Current value: feature_fraction=0.8[LightGBM] [Warning] bagging_fraction is set=0.8, subsample=1.0 will be ignored. Current value: bagging_fraction=0.8Fold: 3, AUC_train: 0.9863, AUC_val: 0.9457, F1-score_train: 0.9853, F1-score_val: 0.9328[LightGBM] [Warning] feature_fraction is set=0.8, colsample_bytree=1.0 will be ignored. Current value: feature_fraction=0.8[LightGBM] [Warning] bagging_fraction is set=0.8, subsample=1.0 will be ignored. Current value: bagging_fraction=0.8Fold: 4, AUC_train: 0.9876, AUC_val: 0.9607, F1-score_train: 0.9860, F1-score_val: 0.9490[LightGBM] [Warning] feature_fraction is set=0.8, colsample_bytree=1.0 will be ignored. Current value: feature_fraction=0.8[LightGBM] [Warning] bagging_fraction is set=0.8, subsample=1.0 will be ignored. Current value: bagging_fraction=0.8Fold: 5, AUC_train: 0.9858, AUC_val: 0.9505, F1-score_train: 0.9841, F1-score_val: 0.9403与LightGBM预测不同的样本数： 724135 -1431761434 1442 1501 -1Name: label, dtype: int64

4.2 随机森林（0.96324）

Ref:
[1] Permutation Importance vs Random Forest Feature Importance (MDI)

from sklearn.ensemble import RandomForestClassifierforest = RandomForestClassifier(max_depth=5, random_state=1234)forest.fit(X_train, Y_train)pred_forest = forest.predict(X_test)result=pd.read_csv('提交示例.csv')result['label']=pred_forestresult.to_csv('result_RandomForest.csv',index=False)feature_importance_forest = pd.Series(forest.feature_importances_, index=X_train.columns).sort_values(ascending=True)plt.figure(figsize=(10, 7), dpi=80)ax = feature_importance_forest.plot.barh()ax.set_title("Random Forest Feature Importances (MDI)")# ax.figure.tight_layout()

## 网格搜索最优参数组合params_test = {'max_depth': range(3, 20, 2),'n_estimators': range(100, 600, 100),'min_samples_leaf': [2, 4, 6]}skf = StratifiedKFold(n_splits=5) gsearch2 = GridSearchCV(estimator=RandomForestClassifier(n_estimators=200, max_depth=5, random_state=random_state), param_grid=params_test, scoring='roc_auc', cv=skf, n_jobs=-1)gsearch2.fit(X_train, Y_train)print(gsearch2.best_params_)print(gsearch2.best_score_)

{'max_depth': 13, 'min_samples_leaf': 2, 'n_estimators': 400}0.9927105570462718

model_forest = RandomForestClassifier(n_estimators=400, max_depth=13, random_state=random_state)result_SKFold_forest = SKFold(X_train, Y_train, X_test, model_forest) result_SKFold_forest.to_csv('result_skfold_RandomForest.csv',index=False)diff_skold_forest = evaluate(result_LightGBM, result_SKFold_forest)

Fold: 1, AUC_train: 0.9935, AUC_val: 0.9516, F1-score_train: 0.9935, F1-score_val: 0.9424Fold: 2, AUC_train: 0.9910, AUC_val: 0.9564, F1-score_train: 0.9909, F1-score_val: 0.9468Fold: 3, AUC_train: 0.9919, AUC_val: 0.9490, F1-score_train: 0.9919, F1-score_val: 0.9396Fold: 4, AUC_train: 0.9919, AUC_val: 0.9647, F1-score_train: 0.9919, F1-score_val: 0.9551Fold: 5, AUC_train: 0.9913, AUC_val: 0.9591, F1-score_train: 0.9912, F1-score_val: 0.9497与LightGBM预测不同的样本数： 118135 -147 -152 -1601641761781851618 -1796 -1Name: label, dtype: int64

4.3 XGBoost (0.95981)

Ref:
[1] XGBoost：在Python中使用XGBoost
[2] Python机器学习笔记：XgBoost算法
[3] python包xgboost安装和简单使用
[4] 深入理解XGBoost，优缺点分析，原理推导及工程实现
[5] XGBoost的原理、公式推导、Python实现和应用
[6] XGBoost官方文档

import xgboost as xgbfrom xgboost import plot_importancefrom sklearn.metrics import roc_auc_scorefrom sklearn.model_selection import StratifiedKFoldimport warningswarnings.filterwarnings('ignore')# 分层k折交叉检验skf = StratifiedKFold(n_splits=5) result_xgb = []fold = 1for train_idx, val_idx in skf.split(X_train, Y_train):train_x = X_train.loc[train_idx]train_y = Y_train.loc[train_idx]val_x = X_train.loc[val_idx]val_y = Y_train.loc[val_idx]d_train = xgb.DMatrix(train_x, train_y)d_val = xgb.DMatrix(val_x, val_y)d_test = xgb.DMatrix(X_test)params = {'max_depth':5,'min_child_weight':1,'num_class':2,'eta': 0.1, #学习率'gamma': 0.1, #后剪枝参数，取值在[0, 1]，越大越保守'seed': 1234,'alpha': 1, #L1正则项的惩罚系数'eval_metric': 'auc'}num_round = 500# # 方式一：采用sklearn接口，采用fit 和 predict# model_xgb = xgb.XGBClassifier()# model_xgb.fit(train_x, train_y, verbose=False) # pred_train = model_xgb.predict(train_x)# pred_val = model_xgb.predict(val_x)# pred_xgb = model_xgb.predict(X_test)# 方式二：采用xgboost原生接口，采用train和predict，方便调参model_xgb = xgb.train(params, d_train, num_round)pred_train = model_xgb.predict(d_train)pred_val = model_xgb.predict(d_val)pred_xgb = model_xgb.predict(d_test)auc_train = roc_auc_score(train_y, pred_train)auc_val = roc_auc_score(val_y, pred_val)f_score_train = f1_score(train_y, pred_train)f_score_val = f1_score(val_y, pred_val)print('Fold: %d, AUC_train: %.4f, AUC_val: %.4f, F1-score_train: %.4f, F1-score_val: %.4f'%(fold, auc_train, auc_val, f_score_train, f_score_val))result_xgb.append(pred_xgb)fold += 1result_xgb = pd.DataFrame(result_xgb).Tprint('result_xgb.shape = ', result_xgb.shape)# 将5次预测结果求平均值result_xgb['average'] = result_xgb.mean(axis=1)# 最终预测结果result_xgb['label'] = result_xgb['average'].apply(lambda x:1 if x>0.5 else 0)# 特征重要性plot_importance(model_xgb)plt.show()# 导出结果result = pd.read_csv('提交示例.csv')result['label'] = result_xgb['label']result.to_csv('result_XGBoost_StratifiedKFold.csv',index=False)diff_xgb = evaluate(result_LightGBM, result_xgb)

Fold: 1, AUC_train: 0.9935, AUC_val: 0.9463, F1-score_train: 0.9929, F1-score_val: 0.9349Fold: 2, AUC_train: 0.9952, AUC_val: 0.9556, F1-score_train: 0.9948, F1-score_val: 0.9456Fold: 3, AUC_train: 0.9952, AUC_val: 0.9518, F1-score_train: 0.9948, F1-score_val: 0.9415Fold: 4, AUC_train: 0.9906, AUC_val: 0.9524, F1-score_train: 0.9893, F1-score_val: 0.9407Fold: 5, AUC_train: 0.9924, AUC_val: 0.9573, F1-score_train: 0.9919, F1-score_val: 0.9482result_xgb.shape = (1000, 5)与LightGBM预测不同的样本数： 621 -124 128 -135 -143 174 -1Name: label, dtype: int64

4.4 CatBoost（0.95854）

CatbBoost 是GBDT算法框架的一种改进实现，其主要创新点有：

支持类别性变量。嵌入了自动将类别型特征处理为数值型特征的创新算法。使用了组合类别特征，丰富特征维度。采用排序提升的方法对抗训练集中的噪声点，从而避免梯度估计的偏差，进而解决预测偏移的问题。采用了完全对称树作为基模型。

Ref:
[1] 深入理解CatBoost
[2] Catboost 一个超级简单实用的boost算法

import catboost as cbfrom sklearn.metrics import roc_auc_scorefrom sklearn.model_selection import StratifiedKFold# 分层k折交叉检验skf = StratifiedKFold(n_splits=5) categorical_features_index = np.where(X_train.dtypes != float)[0]print(X_train.columns[categorical_features_index])result_cat = []fold = 1for train_idx, val_idx in skf.split(X_train, Y_train):train_x = X_train.loc[train_idx]train_y = Y_train.loc[train_idx]val_x = X_train.loc[val_idx]val_y = Y_train.loc[val_idx]model_catboost = cb.CatBoostClassifier(eval_metric='AUC', cat_features=categorical_features_index, depth=6, n_estimators=400, learning_rate=0.5, verbose=False)model_catboost.fit(train_x, train_y, eval_set=(val_x, val_y), plot=False)pred_train = model_catboost.predict(train_x)pred_val = model_catboost.predict(val_x)auc_train = roc_auc_score(train_y, pred_train)auc_val = roc_auc_score(val_y, pred_val)f_score_train = f1_score(train_y, pred_train)f_score_val = f1_score(val_y, pred_val)print('Fold: %d, AUC_train: %.4f, AUC_val: %.4f, F1-score_train: %.4f, F1-score_val: %.4f'%(fold, auc_train, auc_val, f_score_train, f_score_val))pred_catboost = model_catboost.predict(X_test)result_cat.append(pred_catboost)fold += 1result_cat = pd.DataFrame(result_cat).Tprint('result_cat.shape = ', result_cat.shape)# 将5次预测结果求平均值result_cat['average'] = result_cat.mean(axis=1)# 最终预测结果result_cat['label'] = result_cat['average'].apply(lambda x:1 if x>0.5 else 0)# 导出结果result = pd.read_csv('提交示例.csv')result['label'] = result_cat['label']result.to_csv('result_CatBoost_StratifiedKFold.csv',index=False)diff_catboost = evaluate(result_LightGBM, result_cat)feature_importance_catboost = model_catboost.feature_importances_plt.figure(figsize=(10,8), dpi=80)plt.barh(col_names, feature_importance_catboost)plt.show()

Index(['性别', '糖尿病家族史', 'BMI', 'DBP'], dtype='object')Fold: 1, AUC_train: 0.9902, AUC_val: 0.9419, F1-score_train: 0.9887, F1-score_val: 0.9305Fold: 2, AUC_train: 0.9775, AUC_val: 0.9608, F1-score_train: 0.9722, F1-score_val: 0.9510Fold: 3, AUC_train: 0.9988, AUC_val: 0.9543, F1-score_train: 0.9984, F1-score_val: 0.9442Fold: 4, AUC_train: 0.9590, AUC_val: 0.9448, F1-score_train: 0.9498, F1-score_val: 0.9298Fold: 5, AUC_train: 0.9651, AUC_val: 0.9484, F1-score_train: 0.9578, F1-score_val: 0.9377result_cat.shape = (1000, 5)与LightGBM预测不同的样本数： 168133 -135 -147 -152 -160164174 -176183 -185189 -1166 -1501 -1618 -1796 -1Name: label, dtype: int64

4.5 AdaBoost（0.96098）

from sklearn.ensemble import AdaBoostClassifierfrom sklearn.tree import DecisionTreeClassifierrandom_state = 1234model_tree = DecisionTreeClassifier(max_depth=5, random_state=random_state)model_adaboost = AdaBoostClassifier(base_estimator=model_tree, n_estimators=200, random_state=random_state)result_adaboost = SKFold(X_train, Y_train, X_test, model_adaboost) result_adaboost.to_csv('result_AdaBoost.csv',index=False)# 评估diff_skold_adaboost = evaluate(result_LightGBM, result_adaboost)# 特征重要性feature_importance_adaboost = model_adaboost.feature_importances_plt.figure(figsize=(10,8), dpi=80)plt.rc('font', size = 18)plt.barh(col_names, feature_importance_adaboost)plt.title('Feature importances computed by AdaBoost')plt.show()

Fold: 1, AUC_train: 1.0000, AUC_val: 0.9458, F1-score_train: 1.0000, F1-score_val: 0.9347Fold: 2, AUC_train: 1.0000, AUC_val: 0.9445, F1-score_train: 1.0000, F1-score_val: 0.9333Fold: 3, AUC_train: 1.0000, AUC_val: 0.9380, F1-score_train: 1.0000, F1-score_val: 0.9263Fold: 4, AUC_train: 1.0000, AUC_val: 0.9488, F1-score_train: 1.0000, F1-score_val: 0.9357Fold: 5, AUC_train: 1.0000, AUC_val: 0.9536, F1-score_train: 1.0000, F1-score_val: 0.9432与LightGBM预测不同的样本数： 1235 -147 -150 -152 -160164176185192 -1941495 1796 -1Name: label, dtype: int64

4.6 集成模型（0.95971）

挑几个表现较好的模型进行集成。

%%timeskf = StratifiedKFold(n_splits=5) categorical_features_index = np.where(X_train.dtypes != float)[0]print('类别型特征： ', X_train.columns[categorical_features_index])cat_features = list(map(lambda x:int(x), categorical_features_index))random_state = 1234fold = 1for train_idx, val_idx in skf.split(X_train, Y_train):train_x = X_train.loc[train_idx]train_y = Y_train.loc[train_idx]val_x = X_train.loc[val_idx]val_y = Y_train.loc[val_idx]d_train = xgb.DMatrix(train_x, train_y)d_val = xgb.DMatrix(val_x, val_y)d_test = xgb.DMatrix(X_test)params_xgb = {'max_depth':5,'min_child_weight':1,'num_class':2,'eta': 0.1, #学习率'gamma': 0.1, #后剪枝参数，取值在[0, 1]，越大越保守'seed': 1234,'alpha': 1, #L1正则项的惩罚系数'eval_metric': 'auc'}num_round = 500early_stopping_rounds = 100model_lightGBM = lgb.LGBMClassifier(boosting_type='gbdt',objective='binary',metrics='auc',learning_rate=0.1, n_estimators=200, num_leaves=10, silent=True,max_depth=7)model_lightGBM.fit(X_train, Y_train)model_forest = RandomForestClassifier(max_depth=13, n_estimators=400, random_state=1234)model_forest.fit(X_train, Y_train)model_tree = DecisionTreeClassifier(max_depth=5, random_state=random_state)model_adaboost = AdaBoostClassifier(base_estimator=model_tree, n_estimators=200, random_state=random_state)model_adaboost.fit(X_train, Y_train)model_xgb = xgb.train(params_xgb, d_train, num_round)model_catboost = cb.CatBoostClassifier(eval_metric='AUC', cat_features=categorical_features_index, depth=6, iterations=400, learning_rate=0.5, verbose=False)model_catboost.fit(train_x, train_y, eval_set=(val_x, val_y), plot=False)print('Fold: %d finished training. '%fold)fold += 1pred_lightGBM = model_lightGBM.predict(test_data)# pred_lightGBM = list(map(lambda x: 1 if x>0.5 else 0, pred_lightGBM)) #调用lightGBM原生接口时使用pred_forest = forest.predict(X_test)pred_adaboost = model_adaboost.predict(X_test)pred_xgb = model_xgb.predict(d_test)pred_catboost = model_catboost.predict(X_test)pred_all = pd.DataFrame({'lightGBM': pred_lightGBM,'RandomForest': pred_forest,'AdaBoost': pred_adaboost,'XGBoost': pred_xgb,'CatBoost': pred_catboost})pred_all['Average'] = pred_all.mean(axis=1)# 最终预测结果pred_all['label'] = pred_all['Average'].apply(lambda x:1 if x>0.5 else 0)# 导出结果result = pd.read_csv('提交示例.csv')result['label'] = pred_all['label']result.to_csv('result_Ensemble.csv',index=False)diff_ensemble = evaluate(result_LightGBM, result)

类别型特征： Index(['性别', '糖尿病家族史', 'BMI', 'DBP'], dtype='object')Custom logger is already specified. Specify more than one logger at same time is not thread safe.Fold: 1 finished training. Fold: 2 finished training. Fold: 3 finished training. Fold: 4 finished training. Fold: 5 finished training. 与LightGBM预测不同的样本数： 1124133 -135 -143152 -1601641761781851796 -1Name: label, dtype: int64Wall time: 1min 31s

4.7 Stacking（0.96577）

Stacking的思想为在初始数据集上训练若干个基学习器，并将这几个基学习器的预测结果作为新的训练集，来训练一个新的学习器，并将其预测结果作为最终输出。

Stacking本质是一种层级结构，第一层有n个基学习器，每个基学习器进行k折交叉训练，把每一折的验证集（validation set）的预测结果输出并拼接在一起，把这n个模型的训练集预测结果作为新的训练集，将这n个模型的测试集预测结果拼接在一起作为新的测试集。

（图源见水印）

Ref:
[1] stacking模型融合
[2] Kaggle上分技巧——单模K折交叉验证训练+多模型融合

model_tree = DecisionTreeClassifier(max_depth=5, random_state=random_state)clfs = [lgb.LGBMClassifier(boosting_type='gbdt',objective='binary',metrics='auc',learning_rate=0.1, n_estimators=200, num_leaves=10, silent=True,max_depth=7),RandomForestClassifier(max_depth=13, n_estimators=400, random_state=1234),AdaBoostClassifier(base_estimator=model_tree, n_estimators=200, random_state=random_state),xgb.XGBClassifier(),cb.CatBoostClassifier(eval_metric='AUC', cat_features=categorical_features_index, depth=6, iterations=400, learning_rate=0.5, verbose=False)]data_train = np.zeros((X_train.shape[0], len(clfs)))data_test = np.zeros((X_test.shape[0], len(clfs)))# 5折stackingn_splits = 5skf = StratifiedKFold(n_splits)# 第一层，训练各个个体学习器for i, clf in enumerate(clfs):# 依次训练各个模型d_test = np.zeros((X_test.shape[0], n_splits)) #存放个体学习器在测试集上的预测输出for fold, (train_idx, val_idx) in enumerate(skf.split(X_train, Y_train)):#5折交叉训练，第j折拿来预测并作为第二层模型的训练集，剩余部分拿来训练模型。train_x = X_train.loc[train_idx]train_y = Y_train.loc[train_idx]val_x = X_train.loc[val_idx]val_y = Y_train.loc[val_idx]clf.fit(train_x, train_y)pred_train = clf.predict(train_x)pred_val = clf.predict(val_x)data_train[val_idx, i] = pred_vald_test[:, fold] = clf.predict(X_test)auc_train = roc_auc_score(train_y, pred_train)auc_val = roc_auc_score(val_y, pred_val)f_score_train = f1_score(train_y, pred_train)f_score_val = f1_score(val_y, pred_val)print('Classifier:%d, Fold: %d, AUC_train: %.4f, AUC_val: %.4f, F1-score_train: %.4f, F1-score_val: %.4f'%(i+1,fold+1, auc_train, auc_val, f_score_train, f_score_val))#对于测试集，直接用这交叉验证训练的每个模型的预测值均值作为新的特征data_test[:, i] = d_test.mean(axis=1)data_train = pd.DataFrame(data_train)data_test = pd.DataFrame(data_test)# 第二层改用高级点的模型，并进行5折交叉训练# model_forest = RandomForestClassifier(max_depth=5, random_state=1234)model_2 = lgb.LGBMClassifier(boosting_type='gbdt',objective='binary',metrics='auc',learning_rate=0.3, n_estimators=200, num_leaves=10, silent=True,max_depth=7)result_stack = SKFold(data_train, Y_train, data_test, model_2)result_stack.to_csv('result_stack.csv', index=False)diff_stack = evaluate(result_LightGBM, result_stack)

Classifier:1, Fold: 1, AUC_train: 0.9894, AUC_val: 0.9465, F1-score_train: 0.9880, F1-score_val: 0.9340Classifier:1, Fold: 2, AUC_train: 0.9883, AUC_val: 0.9554, F1-score_train: 0.9870, F1-score_val: 0.9465Classifier:1, Fold: 3, AUC_train: 0.9886, AUC_val: 0.9559, F1-score_train: 0.9873, F1-score_val: 0.9467Classifier:1, Fold: 4, AUC_train: 0.9897, AUC_val: 0.9417, F1-score_train: 0.9883, F1-score_val: 0.9268Classifier:1, Fold: 5, AUC_train: 0.9887, AUC_val: 0.9542, F1-score_train: 0.9879, F1-score_val: 0.9453Classifier:2, Fold: 1, AUC_train: 0.9926, AUC_val: 0.9489, F1-score_train: 0.9925, F1-score_val: 0.9377Classifier:2, Fold: 2, AUC_train: 0.9929, AUC_val: 0.9577, F1-score_train: 0.9928, F1-score_val: 0.9482Classifier:2, Fold: 3, AUC_train: 0.9923, AUC_val: 0.9562, F1-score_train: 0.9922, F1-score_val: 0.9478Classifier:2, Fold: 4, AUC_train: 0.9929, AUC_val: 0.9569, F1-score_train: 0.9928, F1-score_val: 0.9470Classifier:2, Fold: 5, AUC_train: 0.9903, AUC_val: 0.9529, F1-score_train: 0.9902, F1-score_val: 0.9439Classifier:3, Fold: 1, AUC_train: 1.0000, AUC_val: 0.9377, F1-score_train: 1.0000, F1-score_val: 0.9253Classifier:3, Fold: 2, AUC_train: 1.0000, AUC_val: 0.9478, F1-score_train: 1.0000, F1-score_val: 0.9354Classifier:3, Fold: 3, AUC_train: 1.0000, AUC_val: 0.9554, F1-score_train: 1.0000, F1-score_val: 0.9465Classifier:3, Fold: 4, AUC_train: 1.0000, AUC_val: 0.9395, F1-score_train: 1.0000, F1-score_val: 0.9269Classifier:3, Fold: 5, AUC_train: 1.0000, AUC_val: 0.9495, F1-score_train: 1.0000, F1-score_val: 0.9399[00:31:58] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.5.0/src/:1115: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.Classifier:4, Fold: 1, AUC_train: 0.9997, AUC_val: 0.9452, F1-score_train: 0.9997, F1-score_val: 0.9326[00:31:59] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.5.0/src/:1115: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.Classifier:4, Fold: 2, AUC_train: 0.9997, AUC_val: 0.9530, F1-score_train: 0.9997, F1-score_val: 0.9429[00:31:59] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.5.0/src/:1115: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.Classifier:4, Fold: 3, AUC_train: 1.0000, AUC_val: 0.9538, F1-score_train: 1.0000, F1-score_val: 0.9441[00:31:59] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.5.0/src/:1115: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.Classifier:4, Fold: 4, AUC_train: 0.9994, AUC_val: 0.9480, F1-score_train: 0.9994, F1-score_val: 0.9345[00:32:00] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.5.0/src/:1115: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.Classifier:4, Fold: 5, AUC_train: 0.9994, AUC_val: 0.9537, F1-score_train: 0.9994, F1-score_val: 0.9452Classifier:5, Fold: 1, AUC_train: 0.9980, AUC_val: 0.9411, F1-score_train: 0.9977, F1-score_val: 0.9293Classifier:5, Fold: 2, AUC_train: 0.9992, AUC_val: 0.9587, F1-score_train: 0.9990, F1-score_val: 0.9485Classifier:5, Fold: 3, AUC_train: 0.9990, AUC_val: 0.9556, F1-score_train: 0.9987, F1-score_val: 0.9456Classifier:5, Fold: 4, AUC_train: 0.9997, AUC_val: 0.9464, F1-score_train: 0.9997, F1-score_val: 0.9321Classifier:5, Fold: 5, AUC_train: 0.9987, AUC_val: 0.9447, F1-score_train: 0.9987, F1-score_val: 0.9326Fold: 1, AUC_train: 0.9582, AUC_val: 0.9455, F1-score_train: 0.9491, F1-score_val: 0.9337Fold: 2, AUC_train: 0.9538, AUC_val: 0.9495, F1-score_train: 0.9449, F1-score_val: 0.9398Fold: 3, AUC_train: 0.9569, AUC_val: 0.9471, F1-score_train: 0.9477, F1-score_val: 0.9361Fold: 4, AUC_train: 0.9546, AUC_val: 0.9556, F1-score_train: 0.9451, F1-score_val: 0.9456Fold: 5, AUC_train: 0.9544, AUC_val: 0.9518, F1-score_train: 0.9463, F1-score_val: 0.9416与LightGBM预测不同的样本数： 140-18123 -128 -133 -135 -147 -152 -160164174 -189 -1796 -1851 -1Name: label, dtype: int64

4.8 归一化数据，pytorch神经网络

先归一化数据，统一量纲。从本节开始，使用基于距离的模型，不再使用树模型。

注意pytorch做二元分类有以下几种实现方式：

Linear 输出维度为1 + sigmoid + BCELoss。Linear 输出维度为1 + BCEWithLogitsLoss。不需要加sigmoid或softmax函数，BCEWithLogitsLoss自带sigmoid作为激活函数。Linear 输出维度为2 + 交叉熵（CrossEntropyLoss)。输出tensor的维度0对应第一个label（即0），维度1对应第二个label（即1）。注意使用交叉熵时，真实标签不能是onehot格式，必须为1维tensor，预测标签必须大于或等于2维，预测标签的每一个维度对应一个标签。

pytorch 中使用神经网络进行多分类时，网络的输出 prediction 是 one hot 格式，但计算交叉熵损失函数时，loss = criterion(prediction, target) 的输入 target 不能是 one hot 格式，直接用数字来表示就行（4 表示 one hot 中的 0 0 0 1）。
所以，自己构建数据集，返回的 target 不需要是 one hot 格式。

Ref:
[1] Pytorch学习笔记(5)——交叉熵报错RuntimeError: 1D target tensor expected, multi-target not supported
[2] PyTorch二分类时BCELoss，CrossEntropyLoss，Sigmoid等的选择和使用
[3] Pytorch实现二分类器
[4] RuntimeError: multi-target not supported at

from sklearn.preprocessing import MinMaxScaler# 归一化scaler = MinMaxScaler()X_train2 = scaler.fit_transform(X_train)X_test2 = scaler.fit_transform(X_test)Y_train2 = Y_train.to_numpy()print('X_train.shape = ', X_train.shape)print('X_train2.shape = ', X_train2.shape)print('Y_train2.shape = ', Y_train2.shape)

X_train.shape = (5070, 10)X_train2.shape = (5070, 10)Y_train2.shape = (5070,)

def Convert(x):# Conver the numeric values into categorical values.y = np.zeros((x.shape[0],))for i in range(len(x)):if x[i, 0] > x[i, 1]:y[i] = 0else:y[i] = 1return y

import torchimport torch.nn as nnimport torch.nn.functional as Fclass NET(nn.Module):def __init__(self, input_dim:int, hidden:int, out_dim:int, activation='relu', dropout=0.2):super(NET, self).__init__()self.input_dim = input_dimself.hidden = hiddenself.out_dim = out_dimself.activation = activationself.Dropout = dropout# 激活函数选择if self.activation == 'relu':mid_act = torch.nn.ReLU()elif self.activation == 'tanh':mid_act = torch.nn.Tanh()elif self.activation == 'sigmoid':mid_act = torch.nn.Sigmoid()elif self.activation == 'LeakyReLU':mid_act = torch.nn.LeakyReLU()elif self.activation == 'ELU':mid_act = torch.nn.ELU()elif self.activation == 'GELU':mid_act = torch.nn.GELU()self.model = nn.Sequential(nn.Linear(self.input_dim, self.hidden),mid_act,nn.Dropout(self.Dropout),nn.Linear(self.hidden, self.hidden),mid_act,nn.Dropout(self.Dropout),nn.Linear(hidden, self.out_dim))def forward(self, x):out = self.model(x)return outdef predict(self, x):# x = torch.tensor(x.to_numpy()) #针对datafram# x = x.to(torch.float32)x = torch.tensor(x).to(torch.float32) #针对ndarrayx = F.softmax(self.model(x))ans = []for t in x:if t[0] > t[1]:ans.append(0)else:ans.append(1)return np.array(ans)

import timefrom torch.utils.data import DataLoader, TensorDatasetclass NN_classifier():def __init__(self, model, crit, l_rate, batch_size, max_epochs, n_splits=5, verbose=True):super(NN_classifier, self).__init__()self.model = model # Neural network model, should be a nn.Module()self.l_rate = l_rateself.batch_size = batch_sizeself.max_epochs = max_epochsself.verbose = verboseself.n_splits = n_splits # the value of k in k-fold validationself.crit = crit # loss functionself.device = 'cpu'def fit(self, X_train, Y_train, X_test):skf = StratifiedKFold(n_splits=self.n_splits)fold = 1pred_test = []for train_idx, val_idx in skf.split(X_train, Y_train):train_x = X_train[train_idx, :]train_y = Y_train[train_idx]val_x = X_train[val_idx, :]val_y = Y_train[val_idx]train_data = TensorDataset(train_x, train_y)train_dataloader = DataLoader(dataset=train_data, batch_size=self.batch_size, shuffle=True)valid_data = TensorDataset(val_x, val_y)validation_dataloader = DataLoader(dataset=valid_data, batch_size=self.batch_size, shuffle=False)model = self.modeloptimizer = torch.optim.Adam(model.parameters(), lr=self.l_rate)scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=50, gamma=0.9) #动态学习率调整for epoch in range(self.max_epochs):start_time = time.time()loss_all = []#------------- Training -----------------model.train()for data in train_dataloader:x, y = datax = x.to(self.device)y = y.to(self.device)optimizer.zero_grad()out = model(x)loss = self.crit(out, y.long())loss.requires_grad_(True) loss.backward()optimizer.step()loss_all.append(loss.item())scheduler.step()end_time = time.time()cost_time = end_time - start_timetrain_loss = np.mean(np.array(loss_all))#------------- Validation -----------------model.eval()loss_all = []with torch.no_grad():for data in validation_dataloader:x, y = datax = x.to(self.device)y = y.to(self.device)output = model(x)loss = self.crit(output, y.long())loss_all.append(loss.item())validation_loss = np.mean(np.array(loss_all))if self.verbose and (epoch+1) % 100 ==0:print('Fold:{:d}, Epoch:{:d}, train_loss: {:.4f}, validation_loss: {:.4f}, cost_time: {:.2f}s'.format(fold, epoch+1, train_loss, validation_loss, cost_time))#------------- Prediction -----------------pred = Convert(model(X_test).detach().numpy())pred_test.append(pred)pred_train = Convert(model(train_x).detach().numpy())pred_val = Convert(model(val_x).detach().numpy())auc_train = roc_auc_score(train_y, pred_train)auc_val = roc_auc_score(val_y, pred_val)f_score_train = f1_score(train_y, pred_train)f_score_val = f1_score(val_y, pred_val)print('Fold: %d, AUC_train: %.4f, AUC_val: %.4f, F1-score_train: %.4f, F1-score_val: %.4f'%(fold, auc_train, auc_val, f_score_train, f_score_val))fold += 1pred_test = pd.DataFrame(pred_test).Tprint('pred_test.shape = ', pred_test.shape)# 将5次预测结果求平均值pred_test['average'] = pred_test.mean(axis=1)#因为竞赛需要你提交最后的预测判断，而模型给出的预测结果是概率，因此我们认为概率>0.5的即该患者有糖尿病，概率<=0.5的没有糖尿病pred_test['label'] = pred_test['average'].apply(lambda x:1 if x>0.5 else 0)## 导出结果result=pd.read_csv('提交示例.csv')result['label']=pred_test['label']return result

# k折交叉验证不断训练同一个模型，集成不同fold（即不同时刻）的模型的预测结果hidden = 64activation = 'relu'# activation = 'tanh'# crit = nn.MSELoss()crit = nn.CrossEntropyLoss()batch_size = 512*2max_epochs = 500l_rate = 1e-3dropout = 0.1n_splits = 5# Convert to tensorX_train_tensor = torch.from_numpy(X_train2).to(torch.float32)X_test_tensor = torch.from_numpy(X_test2).to(torch.float32)Y_train_tensor = torch.from_numpy(Y_train2).to(torch.float32)model_NN = NET(X_train2.shape[1], hidden, out_dim=2, activation=activation)classifier_NN = NN_classifier(model_NN, crit=crit, batch_size=batch_size, l_rate=l_rate, max_epochs=max_epochs, n_splits=n_splits)result_SKFold_NN = NN_classifier.fit(classifier_NN,X_train_tensor, Y_train_tensor, X_test_tensor,)c = result_LightGBM['label'] - result_SKFold_NN['label']count = 0for i in c:if i != 0:count += 1print('与LightGBM预测不同的样本数： ', count)print(c[c!=0])

Fold:1, Epoch:100, train_loss: 0.3503, validation_loss: 0.3196, cost_time: 0.04sFold:1, Epoch:200, train_loss: 0.2785, validation_loss: 0.2505, cost_time: 0.04sFold:1, Epoch:300, train_loss: 0.2332, validation_loss: 0.2257, cost_time: 0.04sFold:1, Epoch:400, train_loss: 0.2098, validation_loss: 0.2130, cost_time: 0.04sFold:1, Epoch:500, train_loss: 0., validation_loss: 0.2066, cost_time: 0.04sFold: 1, AUC_train: 0.9363, AUC_val: 0.9133, F1-score_train: 0.9254, F1-score_val: 0.8956Fold:2, Epoch:100, train_loss: 0.1813, validation_loss: 0.1555, cost_time: 0.04sFold:2, Epoch:200, train_loss: 0.1671, validation_loss: 0.1449, cost_time: 0.04sFold:2, Epoch:300, train_loss: 0.1540, validation_loss: 0.1444, cost_time: 0.04sFold:2, Epoch:400, train_loss: 0.1513, validation_loss: 0.1384, cost_time: 0.04sFold:2, Epoch:500, train_loss: 0.1426, validation_loss: 0.1379, cost_time: 0.04sFold: 2, AUC_train: 0.9496, AUC_val: 0.9395, F1-score_train: 0.9401, F1-score_val: 0.9269Fold:3, Epoch:100, train_loss: 0.1356, validation_loss: 0.1419, cost_time: 0.04sFold:3, Epoch:200, train_loss: 0.1219, validation_loss: 0.1384, cost_time: 0.04sFold:3, Epoch:300, train_loss: 0.1206, validation_loss: 0.1357, cost_time: 0.04sFold:3, Epoch:400, train_loss: 0.1152, validation_loss: 0.1400, cost_time: 0.04sFold:3, Epoch:500, train_loss: 0.1113, validation_loss: 0.1395, cost_time: 0.04sFold: 3, AUC_train: 0.9625, AUC_val: 0.9550, F1-score_train: 0.9535, F1-score_val: 0.9434Fold:4, Epoch:100, train_loss: 0.1075, validation_loss: 0.1185, cost_time: 0.04sFold:4, Epoch:200, train_loss: 0.1155, validation_loss: 0.1250, cost_time: 0.04sFold:4, Epoch:300, train_loss: 0.1081, validation_loss: 0.1238, cost_time: 0.04sFold:4, Epoch:400, train_loss: 0.1056, validation_loss: 0.1283, cost_time: 0.04sFold:4, Epoch:500, train_loss: 0.0957, validation_loss: 0.1289, cost_time: 0.04sFold: 4, AUC_train: 0.9702, AUC_val: 0.9518, F1-score_train: 0.9629, F1-score_val: 0.9386Fold:5, Epoch:100, train_loss: 0.1064, validation_loss: 0.0951, cost_time: 0.04sFold:5, Epoch:200, train_loss: 0.0983, validation_loss: 0.0978, cost_time: 0.04sFold:5, Epoch:300, train_loss: 0.1028, validation_loss: 0.1055, cost_time: 0.05sFold:5, Epoch:400, train_loss: 0.0954, validation_loss: 0.1065, cost_time: 0.04sFold:5, Epoch:500, train_loss: 0.0935, validation_loss: 0.1073, cost_time: 0.04sFold: 5, AUC_train: 0.9709, AUC_val: 0.9547, F1-score_train: 0.9647, F1-score_val: 0.9455pred_test.shape = (1000, 5)与LightGBM预测不同的样本数： 4230-12-14-18116 -1..985 -1987 -1994 -1995 -1999 -1Name: label, Length: 423, dtype: int64

k折交叉验证不断训练同一个模型，虽然模型最终的表现结果还可以（F1-score_val上去了），但集成各fold（各时期）的模型的结果表现依然糟糕，与baseline —— lightGBM相差甚远，都不用提交就知道分数会很低（0.63左右）了。

说明这网络模型不行啊！

可能原因：

集成时被早期表现较差的模型所拖累；模型本身对表格数据拟合能力不够；模型过拟合；

# 不用k折交叉验证，一个模型用到底hidden = 64activation = 'tanh'# activation = 'tanh'crit = nn.CrossEntropyLoss()batch_size = 128max_epochs = 2000l_rate = 5e-3dropout = 0.1n_splits = 5model_NN2 = NET(X_train2.shape[1], hidden, out_dim=2, activation=activation)classifier_NN2 = NN(model_NN2, crit=crit, batch_size=batch_size, l_rate=l_rate, max_epochs=max_epochs)# result_SKFold_NN = SKFold(pd.DataFrame(X_train2), pd.DataFrame(Y_train2),# pd.DataFrame(X_test2), classifier_NN2, n_splits=5)classifier_NN2.fit(X_train2, Y_train2)result_NN = classifier_NN2.predict(X_test2)# c = result_LightGBM['label'] - result_SKFold_NN['label']c = result_LightGBM['label'] - result_NNcount = 0for i in c:if i != 0:count += 1print('与LightGBM预测不同的样本数： ', count)print(c[c!=0])

Epoch:100, train_loss: 0.2839, cost_time: 0.09sEpoch:200, train_loss: 0.2275, cost_time: 0.09sEpoch:300, train_loss: 0.2072, cost_time: 0.09sEpoch:400, train_loss: 0.2097, cost_time: 0.09sEpoch:500, train_loss: 0.1854, cost_time: 0.09sEpoch:600, train_loss: 0.1906, cost_time: 0.09sEpoch:700, train_loss: 0.1772, cost_time: 0.09sEpoch:800, train_loss: 0.1749, cost_time: 0.09sEpoch:900, train_loss: 0.1726, cost_time: 0.08sEpoch:1000, train_loss: 0.1707, cost_time: 0.08sEpoch:1100, train_loss: 0.1692, cost_time: 0.09sEpoch:1200, train_loss: 0.1575, cost_time: 0.09sEpoch:1300, train_loss: 0.1615, cost_time: 0.09sEpoch:1400, train_loss: 0.1548, cost_time: 0.09sEpoch:1500, train_loss: 0.1611, cost_time: 0.09sEpoch:1600, train_loss: 0.1617, cost_time: 0.09sEpoch:1700, train_loss: 0.1535, cost_time: 0.09sEpoch:1800, train_loss: 0.1650, cost_time: 0.09sEpoch:1900, train_loss: 0.1609, cost_time: 0.08sEpoch:2000, train_loss: 0.1631, cost_time: 0.09s与LightGBM预测不同的样本数： 4242-1.04-1.06-1.07-1.081.0... 986 1.0988 -1.0995 -1.0997 -1.0999 -1.0Name: label, Length: 424, dtype: float64

上面这个结果说明这个神经网络模型学习遇到了瓶颈，很难再提升了。要么改模型，要么改训练方法（k折交叉训练重复训练同一个模型有提升，但提升有限，表现依然扑街），要么改数据。

4.9 SVM

表现不行。

Ref:
[1] Python3《机器学习实战》学习笔记（八）：支持向量机原理篇之手撕线性SVM

from sklearn.svm import SVCmodel_SVM = SVC(C=10) #C越大，对误分类的惩罚越大。result_SKFold_SVM= SKFold(pd.DataFrame(X_train2), pd.DataFrame(Y_train2),pd.DataFrame(X_test2), model_SVM, n_splits=5)diff_SVM = evaluate(result_LightGBM, result_SKFold_SVM)

Fold: 1, AUC_train: 0.9001, AUC_val: 0.8614, F1-score_train: 0.8819, F1-score_val: 0.8311Fold: 2, AUC_train: 0.8987, AUC_val: 0.8737, F1-score_train: 0.8797, F1-score_val: 0.8489Fold: 3, AUC_train: 0.8942, AUC_val: 0.8916, F1-score_train: 0.8744, F1-score_val: 0.8711Fold: 4, AUC_train: 0.8901, AUC_val: 0.8840, F1-score_train: 0.8688, F1-score_val: 0.8614Fold: 5, AUC_train: 0.8945, AUC_val: 0.8836, F1-score_train: 0.8747, F1-score_val: 0.8602与LightGBM预测不同的样本数： 2712-18116 -123 -133 -1..983 -1985 -1993 1995 -1999 -1Name: label, Length: 271, dtype: int64

4.10 sklearn神经网络

from sklearn.neural_network import MLPClassifiermodel_MLP = MLPClassifier(hidden_layer_sizes=128, activation='relu')result_SKFold_MLP = SKFold(pd.DataFrame(X_train2), pd.DataFrame(Y_train2),pd.DataFrame(X_test2), model_MLP, n_splits=5)c = result_LightGBM['label'] - result_SKFold_MLP['label']count = 0for i in c:if i != 0:count += 1print('与LightGBM预测不同的样本数： ', count)print(c[c!=0])

Fold: 1, AUC_train: 0.8780, AUC_val: 0.8394, F1-score_train: 0.8550, F1-score_val: 0.8045Fold: 2, AUC_train: 0.8789, AUC_val: 0.8746, F1-score_train: 0.8558, F1-score_val: 0.8512Fold: 3, AUC_train: 0.8838, AUC_val: 0.8828, F1-score_train: 0.8617, F1-score_val: 0.8599Fold: 4, AUC_train: 0.8869, AUC_val: 0.8959, F1-score_train: 0.8637, F1-score_val: 0.8747Fold: 5, AUC_train: 0.8867, AUC_val: 0.8820, F1-score_train: 0.8650, F1-score_val: 0.8579与LightGBM预测不同的样本数： 2382-18116 -123 -133 -1..979 -1983 -1993 1995 -1999 -1Name: label, Length: 238, dtype: int64

神经网络在这个任务上的表现扑街了。。