Keras 处理 不平衡的数据的分类问题 imbalance data 或者 highly skewed data

处理不平衡的数据集的时候，可以使用对数据加权来提高数量较小类的被选中的概率，具体方式如下

fit(self, x, y, batch_size=32, nb_epoch=10, verbose=1, callbacks=[], validation_split=0.0, validation_data=None, shuffle=True, class_weight=None, sample_weight=None)

class_weight：字典，将不同的类别映射为不同的权值，该参数用来在训练过程中调整损失函数（只能用于训练）

sample_weight：权值的numpy array，用于在训练时调整损失函数（仅用于训练）。可以传递一个1D的与样本等长的向量用于对样本进行1对1的加权，或者在面对时序数据时，传递一个的形式为（samples，sequence_length）的矩阵来为每个时间步上的样本赋不同的权。这种情况下请确定在编译模型时添加了sample_weight_mode=’temporal’。

具体使用可以如下：

设置不同累的权值，如下：类0，权值1；类1，权值50

cw = {0: 1, 1: 50}

训练模型

model.fit(x_train, y_train,batch_size=batch_size,epochs=epochs,verbose=1,callbacks=cbks,validation_data=(x_test, y_test), shuffle=True,class_weight=cw)

如果仅仅是类不平衡，则使用class_weight，sample_weights则是类内样本之间还不平衡的时候使用。

class_weight affects the relative weight of each class in the calculation of the objective function.

sample_weights, as the name suggests, allows further control of the relative weight of samples that belong to the same class.

Class weights are useful when training on highly skewed data sets; for example, a classifier to detect fraudulent transactions.

Sample weights are useful when you don't have equal confidence in the samples in your batch. A common example is performing regression on measurements with variable uncertainty.

https://datascience.stackexchange.com/questions/13490/how-to-set-class-weights-for-imbalanced-classes-in-keras

http://blog.csdn.net/lk7688535/article/details/52875046

https://stackoverflow.com/questions/38891390/keras-lstm-with-class-weights

https://stackoverflow.com/questions/43459317/keras-class-weight-vs-sample-weights-in-the-fit-generator

https://stackoverflow.com/questions/41648129/balancing-an-imbalanced-dataset-with-keras-image-generator

https://stackoverflow.com/questions/41815354/keras-flow-from-directory-over-or-undersample-a-class

http://www.ijetae.com/files/Volume2Issue4/IJETAE_0412_07.pdf

http://machinelearningmastery.com/tactics-to-combat-imbalanced-classes-in-your-machine-learning-dataset/

https://stackoverflow.com/questions/44666910/keras-image-preprocessing-unbalanced-data

http://blog.csdn.net/u011401509/article/details/52625014

https://www.analyticsvidhya.com/blog/2016/09/this-machine-learning-project-on-imbalanced-data-can-add-value-to-your-resume/