analysis.py 文件源码

def kmeans_classify(A, means, metric):
    # set up the lists to return
    data_classes = []
    data_metrics = []

    # set up the distance to be the max number possible
    dist = sys.maxint
    for v in A:  # for every data vector
        index = 0
        for i in range(len(means.tolist())):
            m = means.tolist()[i]
            norm_matrix = np.vstack((v, m))
            if norms.pdist(norm_matrix, metric)[0] < dist:
                dist = norms.pdist(norm_matrix, metric)[0]
                index = i

        data_classes.append([index])
        data_metrics.append([dist])
        dist = sys.maxint

    return np.matrix(data_classes), np.matrix(data_metrics)