修复RandomForestClassifier调用predict_proba缺失X参数报错

使用RandomForestClassifier构建二分类器，建模前先基于单特征AUC得分完成特征选择，后续调用封装的模型训练函数获取AUC指标时运行失败，暂未提供配套数据集。

原始代码 #

依赖导入与数据处理流程 #

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score, roc_auc_score
from sklearn.feature_selection import VarianceThreshold

df_process_label1 = 'AAA'
X = df_process.iloc[:,200:500]
y = df_process[df_process_label1].values

import sklearn
from sklearn.model_selection import train_test_split

X_train, X_test, y_train, y_test = train_test_split(X,y, test_size = 0.2, random_state = 0)

constant_filter = VarianceThreshold(threshold = 0.01)
constant_filter.fit(X_train)
X_train_filter = constant_filter.transform(X_train)
X_test_filter = constant_filter.transform(X_test)


roc_auc = []
for features in X_train.columns:
    clf = RandomForestClassifier(n_estimators = 100, random_state=0)
    clf.fit(X_train[features].to_frame(), y_train)
    y_pred = clf.predict(X_test[features].to_frame())
    roc_auc.append(roc_auc_score(y_test, y_pred))


roc_values = pd.Series(roc_auc)
roc_values.index = X_train.columns
roc_values.sort_values(ascending = False, inplace =True)


sel = roc_values[roc_values>0.5]
sel


X_train_roc = X_train[sel.index]
X_test_roc = X_test[sel.index]

def run_randomForest(X_train, X_test, y_train, y_test):
    clf = RandomForestClassifier(n_estimators=100, random_state=0, n_jobs=1)
    clf.fit(X_train, y_train)
    y_pred1 = clf.predict(X_test)
    print('Accuracy on test set: ', accuracy_score(y_test, y_pred))
    print(roc_auc_score(y_test, RandomForestClassifier.predict_proba(X_test)[:,1]))

函数调用代码 #

%time
run_randomForest(X_train_roc, X_test_roc, y_train, y_test)

运行报错信息 #

TypeError: predict_proba() missing 1 required positional argument: 'X'

代码存在两处明确错误，直接导致报错和结果异常：

• predict_proba调用方式错误：直接通过类名RandomForestClassifier调用实例方法，没有绑定训练完成的模型实例clf，方法无法定位要使用的训练好的模型，因此抛出参数缺失错误。正确调用方式为clf.predict_proba(X_test)。
• 准确率计算引用变量错误：函数内当前模型的测试集预测结果被赋值给y_pred1，但计算准确率时传入的是单特征筛选循环中定义的全局变量y_pred，打印的准确率和当前训练的模型无关联，需要统一变量名。
  另外前期方差过滤生成的X_train_filter、X_test_filter后续未被使用，属于冗余代码，不影响运行可自行清理。

修复后的函数代码 #

def run_randomForest(X_train, X_test, y_train, y_test):
    clf = RandomForestClassifier(n_estimators=100, random_state=0, n_jobs=1)
    clf.fit(X_train, y_train)
    y_pred = clf.predict(X_test)
    y_pred_proba = clf.predict_proba(X_test)[:, 1]
    print('Accuracy on test set: ', accuracy_score(y_test, y_pred))
    print('AUC on test set: ', roc_auc_score(y_test, y_pred_proba))

替换原有函数后重新运行，即可正常输出测试集准确率和AUC指标。

内容的提问来源于stack exchange，提问作者Shu