Custom MCC Loss Function in MXNet (Matthews correlation coefficient)

Question

I'm trying to write a MCC Loss Function. I draft out the layout but it has to be done in a Matrix manipulation form as shown in the Reference 1 (This is not a school HW). So the following code is a pseudo code of what I'm trying to do.

class MCCLoss(Loss):
    def __init__(self, weight=None, batch_axis=0, **kwargs):
        super(MCCLoss, self).__init__(weight, batch_axis, **kwargs)

    @staticmethod
    def compute_confusion_matrix_values(y_true, y_pred):
        tp = 0
        fp = 0
        tn = 0
        fn = 0

        for i in range(len(y_pred)):
            if y_true[i] == y_pred[i] == 1:
                tp += 1
            if y_pred[i] == 1 and y_true[i] != y_pred[i]:
                fp += 1
            if y_true[i] == y_pred[i] == 0:
                tn += 1
            if y_pred[i] == 0 and y_true[i] != y_pred[i]:
                fn += 1

        return tp, fp, tn, fn

    @staticmethod
    def matthews_corrcoef(F, tp, fp, tn, fn):
        # https://stackoverflow.com/a/56875660/992687
        x = (tp + fp) * (tp + fn) * (tn + fp) * (tn + fn)
        epsilon = np.finfo(np.float64).eps
        return ((tp * tn) - (fp * fn)) / F.sqrt(x + epsilon)

    def hybrid_forward(self, F, y_pred, y_true, sample_weight=None):
        tp, fp, tn, fn = self.compute_confusion_matrix_values(y_true, y_pred)
        loss = 1 - self.matthews_corrcoef(F, tp, fp, tn, fn)
        return loss

I found some resources that are quite useful, especially an example implementation using Keras in the following Reference 1 link .

I'm not sure if I can use MakeLoss at Reference 2 to simplify the whole thing.

Reference:

Multiple Classification for MCC (Matthews Correlation Coefficient implementation) in Keras https://github.com/vlainic/matthews-correlation-coefficient/blob/master/multi_mcc_loss.py
Custom Loss Function by using MakeLoss http://beta.mxnet.io/r/api/mx.symbol.MakeLoss.html https://blog.csdn.net/u013381011/article/details/79141680
MXNet MCC metric https://github.com/apache/incubator-mxnet/blob/56e79853ad5cf98baf84454eb595c7658bef6ee6/python/mxnet/metric.py#L838

Would any tech expert help me to implement this? Need a good hands

Much appreciatedMuch appreciated

Custom MCC Loss Function in MXNet (Matthews correlation coefficient)

0 Answers0