methods_kharita.py 文件源码

def splitclusters(datapointwts,seeds,theta):
    std = {}; seeds1 = []; seedweight = [];
    cluster, p2cluster = point2cluster(datapointwts, seeds,theta);
    for cl in cluster:
        mang = seeds[cl][-1];
        if len(cluster[cl]) > 10:
            std[cl] = np.percentile([angledist(xx[2], mang) for xx in cluster[cl]], 90)
            clockwise = [xx for xx in cluster[cl] if greaterthanangle(xx[2], mang)];
            if std[cl]>20 and len(clockwise)>0 and len(clockwise)<len(cluster[cl]):
                seeds1.append(avgpoint(clockwise))
                seeds1.append(avgpoint([xx for xx in cluster[cl] if not greaterthanangle(xx[2], mang)]))
                seedweight.append(len(clockwise))
                seedweight.append(len(cluster[cl]) -len(clockwise))
            else:
                seeds1.append(seeds[cl]); seedweight.append(len(cluster[cl]))
        else:
            seeds1.append(seeds[cl]); seedweight.append(len(cluster[cl]))
    return seeds1, seedweight