ML predict function returning wrong value predict(dists, labels, k=1)

Here is the predict function i wrote for Machine learning

```        def predict(dists, labels, k=1):          """ Given a shape-(M, N) array of distances between M-unlabeled                   and N-labeled images, and N labels, predict a label for each               of the M images based on its k-nearest neighbors.                dists : numpy.ndarray                  `dists.shape` must be (M, N) where M is the number of                  examples you wish to predict labels for, and N is                   the number of labeled images used in the prediction                        labels : numpy.ndarray                  A shape-(N,) array of class-IDs, of labels for the                   N images.                    Rtns: y_pred : numpy.array`                  A shape-(M,) array of class-IDs, as predicted by the                        k-nearest neighbors.          """          num_test = dists.shape[0]          y_pred = np.zeros(num_test)          for i in range(num_test):              close_y = []              k_nearest_idxs = np.argsort(dists[i])[:k]              close_y  = labels[k_nearest_idxs]              y_pred[i] = np.argmax(np.bincount(close_y))          return y_pred.astype(np.int64)  ```  

This is how I call this function with dists and labels as below:
d=np.array([[0., 1., 1., 1.]])
l=np.array([0, 1, 1, 0])
predict(d, l, 3)

Expected return value for this function is:  array([0]) but my  function returns array([1])  dists and lables are numpy arrays and k=3(default is 1)  

It fails for other data as well but this is the first of the failure i seem to see.

Any suggestions.. Thx in Advance.  

https://stackoverflow.com/questions/66913952/ml-predict-function-returning-wrong-value-predictdists-labels-k-1 April 02, 2021 at 11:07AM