使用pyod包进行离群点检测将各个方法ROC图绘制在一张图上

# 导包
from pyod.models.lof import LOF
from pyod.models.knn import KNN
from pyod.models.iforest import IForest

import pandas as pd
import matplotlib.pyplot as plt
from sklearn.metrics import roc_curve, auc, roc_auc_score

# 导入数据
df = pd.read_csv('Pima.csv')
X = df.drop(['label'], axis=1)
y = df['label']

#########创建KNN分类器##############
clf_knn = 'KNN'
clfKnn = KNN()
clfKnn.fit(X)

y_pred_knn = clfKnn.predict(X)
y_pred_scores_knn = clfKnn.decision_function(X)

fpr_knn, tpr_knn, threshold_knn = roc_curve(y, y_pred_scores_knn)
roc_auc_knn = auc(fpr_knn, tpr_knn)
#########创建LOF分类器##############
clf_lof = 'LOF'
clfLof = LOF()
clfLof.fit(X)

y_pred_lof = clfLof.predict(X)
y_pred_scores_lof = clfLof.decision_function(X)

fpr_lof, tpr_lof, threshold_lof = roc_curve(y, y_pred_scores_lof)
roc_auc_lof = auc(fpr_lof, tpr_lof)

#########创建IForest分类器##############
clf_iforest = 'IForest '
clfIForest = IForest()
clfIForest.fit(X)

y_pred_iforest = clfIForest.predict(X)
y_pred_scores_iforest = clfIForest.decision_function(X)

fpr_iforest, tpr_iforest, threshold_iforest = roc_curve(y, y_pred_scores_iforest)
roc_auc_iforest = auc(fpr_iforest, tpr_iforest)

plt.figure()
plt.title('Result Analysis')
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')
plt.plot(fpr_knn, tpr_knn, color='green', label='KNN (area = %0.2f)' % roc_auc_knn)
plt.plot(fpr_lof, tpr_lof, color='blue', label='LOF(area = %0.2f)' % roc_auc_lof)
plt.plot(fpr_iforest, tpr_iforest, color='red', label='IForest(area = %0.2f)' % roc_auc_iforest)
plt.legend(loc="lower right")
plt.show()

结果：