6f32626effedfaeba3c51debb39d556302922728,Python/phate/phate.py,,calculate_landmark_operator,#,142

Before Change

        // spectral clustering
        if verbose:
            print("Calculating SVD...")
            start = time.time()
        U, S, _ = randomized_svd(diff_op,
                                 n_components=n_svd,
                                 random_state=random_state)
        if verbose:
            print("Calculated SVD in {:.2f} seconds".format(
                time.time() - start))
            start = time.time()
            print("Calculating Kmeans...")
        kmeans = MiniBatchKMeans(n_landmark,
                                 init_size=3 * n_landmark,
                                 batch_size=10000,
                                 random_state=random_state)
        clusters = kmeans.fit_predict(np.matmul(U, np.diagflat(S)))
        landmarks = np.unique(clusters)
        if verbose:
            print("Calculated Kmeans in {:.2f} seconds".format(
                time.time() - start))

        // transition matrices
        if is_sparse:
            pmn = sparse.vstack(
                [sparse.csr_matrix(kernel[clusters == i, :].sum(
                    axis=0)) for i in landmarks])

After Change

                                 n_components=n_svd,
                                 random_state=random_state)
        log_complete("SVD")
        log_start("KMeans")
        kmeans = MiniBatchKMeans(n_landmark,
                                 init_size=3 * n_landmark,
                                 batch_size=10000,
                                 random_state=random_state)

Instances