train_mlp.py 文件源码

def cluster():
    df_train = pd.read_csv(Train_CSV_Path,header=0)
    destination = []
    for i in range(len(df_train)):
        destination.append(list(eval(df_train['DESTINATION'][i])))

    destination = np.array(destination)
    bw = estimate_bandwidth(
            destination,
            quantile = 0.1,
            n_samples = 1000
            )
    ms = MeanShift(
            bandwidth = bw,
            bin_seeding = True,
            min_bin_freq = 5
            )
    ms.fit(destination)
    cluster_centers = ms.cluster_centers_
    with h5py.File('cluster.h5','w') as f:
        f.create_dataset('cluster',data = cluster_centers)
    return cluster_centers