在DBSCAN

Question

我有一组要点[s1,s2,........,s27] 然后我计算点之间的相似度，然后计算csv文件'数据集'中的点之间的距离:( si和sj之间的距离= 1-相似度）和相似度=（（类受点影响i∩类受点影响j）/（受点影响的类i∪受点j影响的类），最后，我得到一个像distancematrix =[[ds1s1,ds1s2,...,ds1s27],[ds2s1,ds2s2,ds2s3,..ds2s27],.....[ds27s1,ds27s2,.....,ds27s27]]的矩阵 我将距离矩阵输入到DBSCAN算法。我的代码如下：

import operator
from functools import reduce
import pandas as pd
import numpy as np
from sklearn.cluster import DBSCAN
import matplotlib.pyplot as plt
from sklearn.decomposition import PCA
data = pd.read_csv("dataset.csv")
points=['s1','s2','s3','s4','s5','s6','s7','s8','s9','s10','s11','512','s13','s14','s15','s16','s17','s18','s19','s20','s21','s22','s23','s24','s25','s26','s27'] 
Class = data ['Class'].values.tolist()
liste =[]
for n in range (0,len(points)):

    a="points_"+points[n]
    a= data[points[n]].values.tolist()
    b="ClassAffectedBy_"+points[n]
    b= []
    for i in range(0,len(a)):
        if a [i] ==1:
            b.append([Class [i]])
    k = 'ClassAffectedByPoints_'+points[n]
    k = []
    for x in b:
        for y in x:
            k.append(y)
   #print(k)
    liste.append(k)

  dist = []

  for a in range (0,len(liste)):
      for b in range (0,len(liste)):

          Intersection=list(set(liste[a])& set(liste[b]))
          Union=list(set().union(liste[a],liste[b]))

          try:
            if a== b  :
              USim =1
              distance=0
              dist.append(distance)

            else:
              USim=len(Intersection)/len( Union)
              distance=1- (len(Intersection)/len( Union))
              dist.append(distance)
          except ZeroDivisionError :
              USim =0
              distance=1
              dist.append(distance)
distancematrix = [dist[x:x+27]for x in range(0,len(dist),27)]

这段代码是要计算点si和sj之间的距离，然后给DBSCAN距离矩阵如下：

#using default values, set metric to 'precomputed'
db = DBSCAN(eps= 0.75, min_samples = 2, metric='precomputed')
#check db
print(db)
db.fit(distancematrix)
#get labels
labels = db.labels_
print(labels)
#get number of clusters
no_clusters = len(set(labels)) - (1 if -1 in labels else 0)
print('No of clusters:', no_clusters)
print('Cluster 0 ', np.nonzero(labels == 0)[0])
print('Cluster 1 : ', np.nonzero(labels == 1)[0])

现在，我想绘制点并将它们聚类。我使用了这段代码：

import matplotlib.pyplot as plt
# Black removed and is used for noise instead.
core_samples_mask = np.zeros_like(db.labels_, dtype=bool)
core_samples_mask[db.core_sample_indices_] = True
unique_labels = set(labels)
colors = [plt.cm.Spectral(each)
      for each in np.linspace(0, 1, len(unique_labels))]
    for k, col in zip(unique_labels, colors):
     if k == -1:
         # Black used for noise.
         col = [0, 0, 0, 1]
         class_member_mask = (labels == k)
         xy = X[class_member_mask & core_samples_mask]
         plt.plot(xy[:, 0], xy[:, 1], 'o', markerfacecolor=tuple(col),
         markeredgecolor='k', markersize=14)
         xy = X[class_member_mask & ~core_samples_mask]
         plt.plot(xy[:, 0], xy[:, 1], 'o', markerfacecolor=tuple(col),
         markeredgecolor='k', markersize=6)
         plt.title('Estimated number of clusters: %d' % no_clusters)
         plt.show()

问题是，当我执行此代码时，它返回的图形不包含所有点（26个点）-仅返回20点。而且这些点没有很好地分组，我不明白这个问题。