How to write a confusion matrix in Python?

Question

I wrote a confusion matrix calculation code in Python:

def conf_mat(prob_arr, input_arr):
        # confusion matrix
        conf_arr = [[0, 0], [0, 0]]

            


        

Answer 1

Only with numpy, we can do as follow considering efficiency:

def confusion_matrix(pred, label, nc=None):
    assert pred.size == label.size
    if nc is None:
        nc = len(unique(label))
        logging.debug("Number of classes assumed to be {}".format(nc))

    confusion = np.zeros([nc, nc])
    # avoid the confusion with `0`
    tran_pred = pred + 1
    for i in xrange(nc):    # current class
        mask = (label == i)
        masked_pred = mask * tran_pred
        cls, counts = unique(masked_pred, return_counts=True)
        # discard the first item
        cls = [cl - 1 for cl in cls][1:]
        counts = counts[1:]
        for cl, count in zip(cls, counts):
            confusion[i, cl] = count
    return confusion

For other features such as plot, mean-IoU, see my repositories.