Question

我试图在多个数据集上绘制多个聚类算法的结果

代码如下

import numpy as np
import matplotlib.pyplot as plt
import matplotlib

from time import time

from sklearn.datasets import load_digits
from sklearn.decomposition import PCA 
from sklearn.preprocessing import scale


from sklearn import cluster, datasets, mixture
from sklearn.neighbors import kneighbors_graph
from sklearn.preprocessing import StandardScaler

np.random.seed(0)


cs_arry=[[0,1,0,1],[1,0,0,1],[0,0,1,1],[1,1,0,1],[1,0,1,1],[0,1,1,1]]
cs_arry=np.array(cs_arry * 20)
colors = np.array([x for x in cs_arry])

#################
n_pts = 1500#####
#################


noisy_circles = datasets.make_circles(n_samples=n_pts, factor=.5, noise=.05)
no_structure = np.random.rand(n_pts, 2), None
my_circles=make_circ(n_samples=n_pts, factor=.35, noise=.025, in_fact=2)
##############################################
blob_centers = [(-10, -10), (-3, 10), (10, 10)] # adjust blob centers for blob data making 
##############################################
blobs=datasets.make_blobs(n_samples=n_pts, n_features=3, centers=blob_centers, cluster_std=3, 
                          center_box=(-20,20), shuffle=True, random_state=0)


clustering_names=['Input','K-means','K-medoids','AP','Spectral','Gaussian','Hierarchical']

nofclusters=3
plt.figure(figsize=(len(clustering_names) * 2 + 3, 9.5))
plt.subplots_adjust(left=.02, right=.98, bottom=.001, top=.96, wspace=0,
                    hspace=0)
data_sets = [noisy_circles, my_circles, blobs, no_structure]
plot_num = 1

for i_dataset, dataset in enumerate(data_sets):
    x, y = dataset
    x = StandardScaler().fit_transform(x)

    bandwidth = cluster.estimate_bandwidth(x, quantile=0.3)
    connectivity = kneighbors_graph(x, n_neighbors=10, include_self=False)
    connectivity = 0.5 * (connectivity + connectivity.T)

#clustering algorithms
    orig_data=cluster.KMeans(n_clusters=1)
    kmeans=cluster.KMeans(n_clusters=nofclusters,init='k-means++')
    Kmedoids=cluster.KMeans(n_clusters=nofclusters,init='random')
    heir = cluster.AgglomerativeClustering(n_clusters=nofclusters, linkage='ward',
                                           connectivity=connectivity)
    spectral = cluster.SpectralClustering(n_clusters=2,
                                          eigen_solver='arpack',
                                          affinity="nearest_neighbors")
    affinity_propagation = cluster.AffinityPropagation(damping=.9,
                                                       preference=-200)
    gmm=mixture.gmm(n_components=nofclusters, covariance_type='diag', random_state=None, 
                    thresh=None, tol=0.001, min_covar=0.001, n_iter=100, n_init=1, params='wmc', 
                    init_params='wmc', verbose=0)

    #####################################
    #####################################

    clustering_algo=[orig_data, kmeans, kmedoids, ap, spectral, gmm, heir]

    #####################################
    #####################################

    for name, algorithm in zip(clustering_names, clustering_algo):

        t0 = time.time()
        algorithm.fit(x)
        t1 = time.time()
        if hasattr(algorithm, 'labels_'):
            y_pred = algorithm.labels_.astype(np.int)
        else:
            y_pred = algorithm.predict(x)

        # plot
        plt.subplot(len(datasets), len(clustering_algo), plot_num)

        if i_dataset == 0:
            plt.title(name, size=18)
        if name=='Input':
            plt.scatter(x[:, 0], x[:, 1], color='k', s=10, marker='o', facecolors='none')
        else: plt.scatter(x[:, 0], x[:, 1], color=colors[y_pred], s=10, marker='o', facecolors='none')


        plt.xlim(-2, 2)
        plt.ylim(-2, 2)
        plt.xticks(())
        plt.yticks(())
        if name!='Input':
            plt.text(.99, .01, ('%.2fs' % (t1 - t0)).lstrip('0'),
                     transform=plt.gca().transAxes, size=15,
                     horizontalalignment='right')
        plot_num += 1

plt.show()

我想知道是否有办法为不同的数据集设置不同数量的聚类。具体来说，我希望圆形数据有2个簇，而其余的组有3个。

我尝试插入

if data_sets==noisy_circles or my_circles: 
     nofclusters=2

或

while data_sets==noisy_circles or my_circles: 
     nofclusters=2

在for循环语句之前和之后没有结果。

感谢任何帮助。

谢谢。

Answer 1

使用字典，这样您就可以轻松配置每个数据集的簇数。您还需要选择许多其他参数，例如DBSCAN中的epsilon和minpts。

显然你需要在里面 for循环，并确保在每次迭代时设置值。否则，下一次迭代可能会得到之前的值，结果会得到错误。

不同的没有。不同数据集的集群

1 个答案: