-
Notifications
You must be signed in to change notification settings - Fork 0
/
DBSCAN.py
115 lines (95 loc) · 3.91 KB
/
DBSCAN.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
import sklearn.cluster as sc
from clustering_metric import clustering_metrics
import numpy as np
from tqdm import tqdm
from utils import load_data
from utils import WL
from run import main
import pandas as pd
import random
import sys
sys.path.append("..")
dataset = 'cora'
adj_mat, node_features, y_test, tx, ty, test_maks, true_labels = load_data(dataset)
print("========================= {} ============================".format(dataset))
list_nmi, list_f1, list_acc = [], [], []
best_acc, best_nmi, best_f1,best_h= 0, 0, 0,'ERROR'
ac_l = [0.01]
h_l = [13]
for h in h_l:
embedding = WL(node_features, adj_mat, h)
model = sc.KMeans(n_clusters=len(list(set(true_labels))))
model = model.fit(embedding)
predict_labels = model.predict(embedding)
# result_list, centerIndex_list = DensityPeak(X=embedding, n_cluster=len(list(set(true_labels)))).autoSearch()
# for predict_labels in result_list:
cm = clustering_metrics(true_labels, predict_labels)
acc,nmi, f1_macro, precision_macro, recall_macro, f1_micro, precision_micro, recall_micro, adjscore = cm.evaluationClusterModelFromLabel(tqdm)
current_acc = acc
current_nmi=nmi
current_f1 = f1_macro
if current_nmi > best_nmi:
best_nmi = current_nmi
best_h = h
best_labels = predict_labels
if current_f1 > best_f1:
best_f1 = current_f1
if current_acc > best_acc:
best_acc = current_acc
print("@h={} best_nmi:{} best_nmi:{} best_f1:{}".format(best_h,best_acc,best_nmi,best_f1))
for train in range(6):
predict_labels =best_labels
adj_mat=np.array(adj_mat[0])
features=[[] for i in range(len(list(set(true_labels))))]
edges =[[] for i in range(len(list(set(true_labels))))]
cluster = [[] for i in range(len(list(set(true_labels))))]
for id in range(len(predict_labels)):
label=predict_labels[id]
edges[label].append(id)
features[label].append(node_features[0][id])
adj=[]
for id in range(len(features)):
i, j = np.ix_(edges[id],edges[id])
adj.append(adj_mat[i,j])
del_li=[]
for i in range(len(list(set(true_labels)))):
score = main(features[i],adj[i],1,2).reshape(-1).tolist()[0]
ac =random.sample(ac_l,1)
n = int(len(features[i])*ac[0])
# 选取list数组元素中最大（最小）的n个值的索引
res = pd.Series(score).sort_values().index[:n].tolist()
res =[edges[i][j] for j in res ]
del_li += res
ll = len(list(set(del_li)))
rr = len(del_li)
features2 = node_features[0]
features2 = np.delete(features2,del_li,axis=0)
adj = np.delete(adj_mat, del_li, axis=1)
adj = np.delete(adj, del_li, axis=0)
node_features2,adj_mat2=[],[]
node_features2.append(features2)
adj_mat2.append(adj)
true_labels2 =np.delete(true_labels, del_li, axis=0)
best_acc, best_nmi, best_f1,= 0, 0, 0,
embedding = WL(node_features2, adj_mat2, best_h)
print(node_features2[0].shape,adj_mat2[0].shape,rr,ll)
model = sc.KMeans(n_clusters=len(list(set(true_labels))))
model = model.fit(embedding)
predict_labels = model.predict(embedding)
cm = clustering_metrics(true_labels2, predict_labels)
acc,nmi, f1_macro, precision_macro, recall_macro, f1_micro, precision_micro, recall_micro, adjscore = cm.evaluationClusterModelFromLabel(tqdm)
current_acc = acc
current_nmi=nmi
current_f1 = f1_macro
if current_nmi > best_nmi:
best_nmi = current_nmi
best_h = h
best_labels = predict_labels
if current_f1 > best_f1:
best_f1 = current_f1
if current_acc > best_acc:
best_acc = current_acc
print("@train={} ac={} best_nmi:{} best_nmi:{} best_f1:{}".format(train,ac[0],best_acc,best_nmi,best_f1))
adj_mat =adj_mat2
node_features = node_features2
true_labels = true_labels2