mlutils.py

import numpy as np
import codecs
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.datasets import make_classification
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import silhouette_samples, silhouette_score
from sklearn.cluster import KMeans

def load_SMS_dataset(path):
    X, y = [], []
    with codecs.open(path, "r", "utf-8") as fd:
        for line in fd:
            parts = line.split("\t")
            X.append(parts[1].strip())
            y.append(parts[0].strip())
    return np.array(X), np.array(y)

def plot_silhouette(n_clusters, X):
    # Kôd preuzet s http://scikit-learn.org/stable/auto_examples/cluster/plot_kmeans_silhouette_analysis.html

    # Create a subplot with 1 row and 2 columns
    fig, (ax1, ax2) = plt.subplots(1, 2)
    fig.set_size_inches(18, 7)

    # The 1st subplot is the silhouette plot
    # The silhouette coefficient can range from -1, 1 but in this example all
    # lie within [-0.1, 1]
    ax1.set_xlim([-0.1, 1])
    # The (n_clusters+1)*10 is for inserting blank space between silhouette
    # plots of individual clusters, to demarcate them clearly.
    ax1.set_ylim([0, len(X) + (n_clusters + 1) * 10])

    # Initialize the clusterer with n_clusters value and a random generator
    # seed of 10 for reproducibility.
    clusterer = KMeans(n_clusters=n_clusters, random_state=10)
    cluster_labels = clusterer.fit_predict(X)

    # The silhouette_score gives the average value for all the samples.
    # This gives a perspective into the density and separation of the formed
    # clusters
    silhouette_avg = silhouette_score(X, cluster_labels)
    print("For n_clusters =", n_clusters,
          "The average silhouette_score is :", silhouette_avg)

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

    y_lower = 10
    for i in range(n_clusters):
        # Aggregate the silhouette scores for samples belonging to
        # cluster i, and sort them
        ith_cluster_silhouette_values = \
            sample_silhouette_values[cluster_labels == i]

        ith_cluster_silhouette_values.sort()

        size_cluster_i = ith_cluster_silhouette_values.shape[0]
        y_upper = y_lower + size_cluster_i

        cmap = plt.cm.get_cmap("Spectral")
        color = cmap(float(i) / n_clusters)
        ax1.fill_betweenx(np.arange(y_lower, y_upper),
                          0, ith_cluster_silhouette_values,
                          facecolor=color, edgecolor=color, alpha=0.7)

        # Label the silhouette plots with their cluster numbers at the middle
        ax1.text(-0.05, y_lower + 0.5 * size_cluster_i, str(i))

        # Compute the new y_lower for next plot
        y_lower = y_upper + 10  # 10 for the 0 samples

    ax1.set_xlabel("Vrijednosti koeficijenta siluete")
    ax1.set_ylabel("Oznaka grupe")

    # The vertical line for average silhouette score of all the values
    ax1.axvline(x=silhouette_avg, color="red", linestyle="--")

    ax1.set_yticks([])  # Clear the yaxis labels / ticks
    ax1.set_xticks([-0.1, 0, 0.2, 0.4, 0.6, 0.8, 1])

    # 2nd Plot showing the actual clusters formed
    cmap = plt.cm.get_cmap("Spectral")
    colors = cmap(cluster_labels.astype(float) / n_clusters)
    ax2.scatter(X[:, 0], X[:, 1], marker='.', s=30, lw=0, alpha=0.7,
                c=colors, edgecolor='k')

    # Labeling the clusters
    centers = clusterer.cluster_centers_
    # Draw white circles at cluster centers
    ax2.scatter(centers[:, 0], centers[:, 1], marker='o',
                c="white", alpha=1, s=200, edgecolor='k')

    for i, c in enumerate(centers):
        ax2.scatter(c[0], c[1], marker='$%d$' % i, alpha=1,
                    s=50, edgecolor='k')

    ax2.set_xlabel(r"$x_1$")
    ax2.set_ylabel(r"$x_2$")

    plt.show()

def plot_2d_clf_problem(X, y, h=None):
    '''
    Plots a two-dimensional labeled dataset (X,y) and, if function h(x) is given, 
    the decision surfaces.
    '''
    assert X.shape[1] == 2, "Dataset is not two-dimensional"
    if h!=None : 
        # Create a mesh to plot in
        r = 0.02  # mesh resolution
        x_min, x_max = X[:, 0].min() - 1, X[:, 0].max() + 1
        y_min, y_max = X[:, 1].min() - 1, X[:, 1].max() + 1
        xx, yy = np.meshgrid(np.arange(x_min, x_max, r),
                             np.arange(y_min, y_max, r))
        XX=np.c_[xx.ravel(), yy.ravel()]
        try:
            Z_test = h(XX)
            if Z_test.shape == ():
                # h returns a scalar when applied to a matrix; map explicitly
                Z = np.array(list(map(h,XX)))
            else :
                Z = Z_test
        except ValueError:
            # can't apply to a matrix; map explicitly
            Z = np.array(list(map(h,XX)))
        # Put the result into a color plot
        Z = Z.reshape(xx.shape)
        plt.contourf(xx, yy, Z, cmap=plt.cm.Pastel1)

    # Plot the dataset
    plt.scatter(X[:,0],X[:,1], c=y, cmap=plt.cm.tab20b, marker='o', s=50);

def plot_2d_svc_problem(X, y, svc=None):
    '''
    Plots a two-dimensional labeled dataset (X,y) and, if SVC object is given, 
    the decision surfaces (with margin as well).
    '''
    assert X.shape[1] == 2, "Dataset is not two-dimensional"
    if svc!=None : 
        # Create a mesh to plot in
        r = 0.03  # mesh resolution
        x_min, x_max = X[:, 0].min() - 1, X[:, 0].max() + 1
        y_min, y_max = X[:, 1].min() - 1, X[:, 1].max() + 1
        xx, yy = np.meshgrid(np.arange(x_min, x_max, r),
                             np.arange(y_min, y_max, r))
        XX=np.c_[xx.ravel(), yy.ravel()]
        Z = np.array([svc_predict(svc, x) for x in XX])
        # Put the result into a color plot
        Z = Z.reshape(xx.shape)
        plt.contourf(xx, yy, Z, cmap=plt.cm.Pastel1)

    # Plot the dataset
    plt.scatter(X[:,0],X[:,1], c=y, cmap=plt.cm.Paired, marker='o', s=50)
    #plt.show()

def svc_predict(svc, x) : 
    h = svc.decision_function([x])
    if h >= -1 and h <= 1:
        return 0.5
    else: 
        return max(-1, min(1, h))

def plot_error_surface(err, c_range=(0,5), g_range=(0,5)):
    c1, c2 = c_range[0], c_range[1]
    g1, g2 = g_range[0], g_range[1]
    plt.xticks(range(0,g2-g1+1,5),range(g1,g2,5)); plt.xlabel("gamma")
    plt.yticks(range(0,c2-c1+1,5),range(c1,c2,5)); plt.ylabel("C")
    p = plt.contour(err);
    plt.imshow(1-err, interpolation='bilinear', origin='lower',cmap=plt.cm.gray)
    plt.clabel(p, inline=1, fontsize=10)
    #plt.show()


def knn_eval(n_instances=100, n_features=2, n_classes=2, n_informative=2, 
             test_size=0.3, k_range=(1, 20), n_trials=100):
    
    train_errors = []
    test_errors = []
    ks = list(range(k_range[0], k_range[1] + 1))

    for i in range(0, n_trials):
        X, y = make_classification(n_instances, n_features, n_classes=n_classes, 
                                   n_informative=n_informative, n_redundant=0, n_clusters_per_class=1)
        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=test_size)
        train = []
        test = []
        for k in ks:
            knn = KNeighborsClassifier(n_neighbors=k)
            knn.fit(X_train, y_train)
            train.append(1 - knn.score(X_train, y_train))
            test.append(1 - knn.score(X_test, y_test))
        train_errors.append(train)
        test_errors.append(test)
        
    train_errors = np.mean(np.array(train_errors), axis=0)
    test_errors = np.mean(np.array(test_errors), axis=0)
    best_k = ks[np.argmin(test_errors)]
    
    return ks, best_k, train_errors, test_errors