/
mnist.py
executable file
·121 lines (102 loc) · 3.17 KB
/
mnist.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
116
117
118
119
120
121
from sklearn.neighbors import LSHForest
from sklearn.datasets.samples_generator import make_blobs
from sklearn.neighbors import NearestNeighbors
import numpy as np
import math
from scipy.interpolate import Rbf
import cv2
import random;
import math
import time
import matplotlib.pyplot as plt
def plot(lsh_acc,knn_acc):
plt.xlabel('K (Nearest neighbors)')
plt.title('Accuracy vs K')
plt.ylabel('Accuracy')
ks=[];
for i in range(len(lsh_acc)):
ks.append(i+1)
print len(lsh_acc)
plt.plot(ks,lsh_acc,'bo--',label="K-NN with LSH")
plt.plot(ks,knn_acc,'go--',label="Exhaustive K-NN")
plt.legend()
plt.show()
# def timing(f):
# def wrap(*args):
# time1 = time.time()
# ret = f(*args)
# time2 = time.time()
# print '%s function took %0.3f ms' % (f.func_name, (time2-time1)*1000.0)
# return ret
# return wrap
def getData():
train = np.genfromtxt('mnist_data/train_data.txt')
test = np.genfromtxt('mnist_data/test_data.txt')
val = np.genfromtxt('mnist_data/val_data.txt')
return (train,test,val)
def applyLSH(lshf,test_data):
indicesLSH = lshf.kneighbors(test_data, n_neighbors=20,return_distance=False)
return indicesLSH
def lshFunct(test,val,n_feat,lshf,train_labels):
# Parse data for separating training labels and dataset
test_data = test[:, :-1]
test_labels = test[:, n_feat - 1]
test_error=[]
training_error=[]
indicesLSH=applyLSH(lshf,test_data)
countarrLSH=[0,0,0,0,0];
for k in range(1,6):
for i in range(len(indicesLSH)):
for j in range(k):
currptr=indicesLSH[i][j];
if(train_labels[currptr]==test_labels[i]):
countarrLSH[k-1]+=1
break
for i in range(len(countarrLSH)):
countarrLSH[i]=((countarrLSH[i])*100)/float(len(indicesLSH))
return countarrLSH
def trainLSH(train,test,val):
n_feat = train[0].size
train_data = train[:, :-1]
train_labels = train[:, n_feat - 1]
val_data = val[:, :-1]
val_labels = val[:, n_feat - 1]
lshf = LSHForest(random_state=42)
lshf.fit(train_data);
countarrLSH=lshFunct(test,val,n_feat,lshf,train_labels)
return countarrLSH
def knnFunct(test,val,n_feat,neigh,train_labels):
# Parse data for separating training labels and dataset
test_labels = test[:, n_feat - 1]
test_data = test[:, :-1]
test_error=[]
training_error=[]
indicesKNN = neigh.kneighbors(test_data, n_neighbors=20,return_distance=False)
countarrKNN=[0,0,0,0,0];
for k in range(1,6):
for i in range(len(indicesKNN)):
for j in range(k):
currptr=indicesKNN[i][j];
if(train_labels[currptr]==test_labels[i]):
countarrKNN[k-1]+=1
break
for i in range(len(countarrKNN)):
countarrKNN[i]=((countarrKNN[i])*100)/float(len(indicesKNN))
return countarrKNN
def trainKNN(train,test,val):
n_feat = train[0].size
train_data = train[:, :-1]
train_labels = train[:, n_feat - 1]
val_data = val[:, :-1]
val_labels = val[:, n_feat - 1]
neigh = NearestNeighbors(20,0.5,metric='cosine',algorithm="brute")
neigh.fit(train_data)
countarrKNN=knnFunct(test,val,n_feat,neigh,train_labels)
return countarrKNN
if __name__ == '__main__':
train,test,val=getData();
accuracyLSH=trainLSH(train,test,val);
accuracyKNN=trainKNN(train,test,val)
print accuracyKNN
print accuracyLSH
plot(accuracyLSH,accuracyKNN)