如何在keras网格搜索中正确实现f1score作为评分指标？

Question

我想用GridSearchCV来调整我的keras神经网络，因为我在数据集中有很高的不平衡性。这是一个二元分类问题。这是我尝试过的

def f1(y_true, y_pred):
   y_pred = K.round(y_pred)
   tp = K.sum(K.cast(y_true*y_pred, 'float'), axis=0)
   # tn = K.sum(K.cast((1-y_true)*(1-y_pred), 'float'), axis=0)
   fp = K.sum(K.cast((1-y_true)*y_pred, 'float'), axis=0)
   fn = K.sum(K.cast(y_true*(1-y_pred), 'float'), axis=0)

   p = tp / (tp + fp + K.epsilon())
   r = tp / (tp + fn + K.epsilon())

   f1 = 2*p*r / (p+r+K.epsilon())
   f1 = tf.where(tf.is_nan(f1), tf.zeros_like(f1), f1)
   return K.mean(f1)

def create_model():
   model = Sequential()
   model.add(Dense(899, activation = 'relu', input_dim = 899))
   model.add(Dense(100, activation = 'relu'))
   model.add(Dense(1, activation = 'sigmoid'))
   model.compile(optimizer= 'adamax', loss = 'binary_crossentropy', metrics = [f1])
   return model


param_grid = dict(batch_size = batch_size, epochs = epochs,class_weight = class_weight)
grid = GridSearchCV(estimator = model, param_grid = param_grid, n_jobs =7, cv = 2, scoring = f1 )

但我得到了错误

^{2}$

看来我也需要创造一个得分手？如何才能正确地做到这一点呢？在

更新：

这似乎解决了这个问题，我认为它是按计划工作的

from sklearn.metrics import f1_score
from sklearn.metrics import make_scorer

f1_scorer = make_scorer(f1_score) 

然后传递给GridsearchCV函数

grid = GridSearchCV(estimator = model, param_grid = param_grid, n_jobs =7, cv = 2, scoring = f1_scorer )