-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathassembly_metrics.py
96 lines (71 loc) · 3.1 KB
/
assembly_metrics.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
import numpy as np
from scipy.spatial.distance import squareform, pdist
import matplotlib.pyplot as plt
from scipy.stats import pearsonr,spearmanr
import torch
import argparse
import utils
from torch_geometric.data import DataLoader
from torch_geometric.datasets import TUDataset
import os
# 计算数据的RDM
def compute_rdm(data_matrix):
# 计算数据矩阵中每对样本之间的欧氏距离
distances = pdist(data_matrix, 'cosine')
# 将距离转换成RDM（代表性差异矩阵）
rdm = squareform(distances)
return rdm
def vis_rdm(rdm_matrix):
# 可视化RDM
plt.imshow(rdm_matrix, cmap='Greys', origin='upper')
plt.colorbar(label='Dissimilarity')
plt.title('Representational Dissimilarity Matrix')
plt.xlabel('Samples')
plt.ylabel('Samples')
plt.savefig('./pics/rdm.jpg')
def compare_rdm(rdm_matrix1,rdm_matrix2):
# 将两个RDMs转换为成对距离矩阵
rdm1 = squareform(rdm_matrix1)
rdm2 = squareform(rdm_matrix2)
# 计算Pearson相关系数和p值
corr_pearson, p_value = pearsonr(rdm1, rdm2)
###print("Pearson Correlation:{:.4f}".format(corr_pearson))
# print("P-value:", p_value)
corr_spear, p_value = spearmanr(rdm1, rdm2)
###print("Spearman Correlation:{:.4f}".format(corr_spear))
# print("P-value:", p_value)
return corr_pearson, corr_spear
def infer(encoder, dataset,args):
data_loader = DataLoader(dataset, batch_size=args.batch_size, shuffle=False)
embeddings_list = []
for i, data in enumerate(data_loader):
data = data.to(args.device)
embeddings = encoder(data).detach().cpu().numpy()
embeddings_list.append(embeddings)
embeddings = np.concatenate(tuple(embeddings_list), axis=0)
return embeddings
def compute_double_rdm(encoder_1,encoder_2,dataset,args):
embeddings_1 = infer(encoder_1, dataset,args)
embeddings_2 = infer(encoder_2, dataset,args)
rdm_matrix_1 = compute_rdm(embeddings_1)
rdm_matrix_2 = compute_rdm(embeddings_2)
corr_pearson, corr_spear = compare_rdm(rdm_matrix_1,rdm_matrix_2)
return corr_pearson, corr_spear
def compute_triple_rdm(encoder_1,encoder_2,encoder_3,dataset,args):
embeddings_1 = infer(encoder_1, dataset,args)
embeddings_2 = infer(encoder_2, dataset,args)
embeddings_3 = infer(encoder_3, dataset,args)
rdm_matrix_1 = compute_rdm(embeddings_1)
rdm_matrix_2 = compute_rdm(embeddings_2)
rdm_matrix_3 = compute_rdm(embeddings_3)
corr_pearson_12, corr_spear_12 = compare_rdm(rdm_matrix_1,rdm_matrix_2)
corr_pearson_13, corr_spear_13 = compare_rdm(rdm_matrix_1,rdm_matrix_3)
corr_pearson_23, corr_spear_23 = compare_rdm(rdm_matrix_2,rdm_matrix_3)
return (corr_pearson_12+corr_pearson_13+corr_pearson_23)/3, (corr_spear_12+corr_spear_13+corr_spear_23)/3 \
, corr_spear_12, corr_spear_13, corr_spear_23
'''
# 创建一个虚拟的神经数据矩阵，每一行代表不同的样本或刺激条件，每一列代表不同的神经单元或特征
neural_data = np.random.rand(10, 20) # 假设有10个样本和20个特征
# 计算RDM
rdm_matrix = compute_rdm(neural_data)
'''