class_weight = {1: (pos_num+neg_num)/2/pos_num, 0: (pos_num+neg_num)/2/neg_num}

This formula calculates the class weights for a binary classification problem where the positive class has a higher number of samples than the negative class.

- `pos_num` is the number of samples in the positive class
- `neg_num` is the number of samples in the negative class
- `(pos_num neg_num)/2` calculates the average number of samples in each class
- `(pos_num neg_num)/2/pos_num` calculates the weight for the positive class as the inverse of the positive class frequency in the dataset
- `(pos_num neg_num)/2/neg_num` calculates the weight for the negative class as the inverse of the negative class frequency in the dataset

The resulting class weights will be used to balance the contribution of each class during training, so that the model can learn to classify both classes equally well.