validation_data=(X_test, Y_test_oh))
# 保存模型
model.save(r'.\model\\' + path + '\\' + str(signal_len) + 's-train_' +
str(k + 1) + '-fold.h5py')
# 预测
prob_trainY = model.predict(X_train)
prob_testY = model.predict(X_test)
pred_trainY = np.argmax(prob_trainY, axis=1)
pred_testY = np.argmax(prob_testY, axis=1)
# 评价指标
Se_tr, Sp_tr, Ber_tr = calc_ber(Y_train, pred_trainY)
Se_te, Sp_te, Ber_te = calc_ber(Y_test, pred_testY)
Auc_tr = calc_auc(Y_train, np.argmax(prob_trainY, axis=1), 1)
Auc_te = calc_auc(Y_test, np.argmax(prob_testY, axis=1), 1)
# 绘制代价函数曲线
train_loss = history.history['loss']
test_loss = history.history['val_loss']
train_auc = history.history['auc']
test_auc = history.history['val_auc']
# 保存重要数据
important_data = {
'train_loss': train_loss,
'test_loss': test_loss,
'Se': Se_te,
'Sp': Sp_te,
prob_trainY = model.predict(X_train)
prob_testY = model.predict(X_test)
# ROC曲线 tpr = Se,fpr = 1-Sp Youden_index = Se+Sp-1 = tpr-fpr(约登指数)
fpr, tpr, thresholds = roc_curve(Y_train,
prob_trainY,
1,
drop_intermediate=False)
Youden_index = tpr - fpr
# 概率转换成预测标签
thr = thresholds[np.where(Youden_index == np.max(Youden_index))]
pred_trainY = (prob_trainY >= thr).astype('int8')
pred_testY = (prob_testY >= thr).astype('int8')
# 评价指标
Se_tr, Sp_tr, Ber_tr = calc_ber(Y_train, pred_trainY)
Se_te, Sp_te, Ber_te = calc_ber(Y_test, pred_testY)
Auc_tr = calc_auc(Y_train, prob_trainY, 1)
Auc_te = calc_auc(Y_test, prob_testY, 1)
# 绘制代价函数曲线
train_loss = history.history['loss']
test_loss = history.history['val_loss']
train_auc = history.history['auc']
test_auc = history.history['val_auc']
# 保存重要数据
important_data = {
'train_loss': train_loss,
'test_loss': test_loss,
'Se': Se_te,
'Sp': Sp_te,
'Auc': Auc_te,