The Silhouette plot for different n_cluster values when clustered using KMeans

silhouette_plot_kmeans.py

# Generating the sample data from make_blobs

X, y = make_blobs(n_samples=100,
                  n_features=2,
                  centers=4,
                  cluster_std=1.0,
                  center_box=(-10.0, 10.0),
                  shuffle=True,
                  random_state=0) # For reproducibility

range_n_clusters = [ 2, 4, 6 ]

fignum = 0

for n_clusters in range_n_clusters:
    fignum += 1
    fig = plt.figure(fignum, figsize = (6, 6))
    plt.grid()
    plt.xlim([-0.1, 1])
    plt.ylim([0, len(X)])

    # Initialize kmeans clusterer with 3 clusters and a random genertor seed of 10
    # for reproducibility

    kmeans_clusterer = KMeans(n_clusters=n_clusters, random_state=10)

    cluster_labels = kmeans_clusterer.fit_predict(X)

    # Compute the silhouette scores for each sample
    sample_silhouette_values = silhouette_samples(X, cluster_labels)


    sorted_clustered_sample_silhouette_values = []

    for i in np.unique(cluster_labels):
        ith_cluster_silhouette_values = sample_silhouette_values[cluster_labels == i]

        # Add the ith_cluster_silhouette_values after sorting them
        ith_cluster_silhouette_values.sort()

        # The 0 sample is to differentiate clearly between the different clusters
        sorted_clustered_sample_silhouette_values += ith_cluster_silhouette_values.tolist() + [0]


    plt.plot(sorted_clustered_sample_silhouette_values, range(len(X)  + n_clusters))
    fig.show()

    fignum += 1
    fig = plt.figure(fignum, figsize = (6, 6))
    plt.grid()
    for k in range(len(X)):
        color = cm.spectral(float(cluster_labels[k]) / n_clusters, 1)
        plt.plot(X[k, 0], X[k, 1], 'o', marker='.', c=color)

    plt.title("The visualization of the clustered data")
    fig.show()