-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathcross_validation.py
190 lines (141 loc) · 6.68 KB
/
cross_validation.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
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
################################## cross_validation.py ##########################
################################## Created By Anirudha Nilegaonkar ##########################
import csv
import numpy as np
import operator
import random
#Function for finding Euclidean Distance
def CalculateEuclideanDistance(input_1, input_2, length):
distance = 0
for i in range(length-1):
distance += (input_1[i] - input_2[i]) **2
Euclidean_distance=distance**(1/2)
return Euclidean_distance
#Function for K finding neighbours
def findNeighbours(final_train_matrix, testInstance, k):
Neighbours_distances = []
for i in range(len(final_train_matrix)):
respective_distance = CalculateEuclideanDistance(testInstance, final_train_matrix[i,1:785],len(testInstance))
#Contains distance values of test Instance w.r.t all train matrix rows
Neighbours_distances.append((final_train_matrix[i],respective_distance))
#Sorting Neighbours_distances with ascending order
Neighbours_distances.sort(key=operator.itemgetter(1))
#Choose first "K" distances
Final_neighbors = []
for i in range(k):
Final_neighbors.append(Neighbours_distances[i][0])
return Final_neighbors
#Function for selecting best neighbour
def findBestNeighbour(find_neighbours):
neighbour_count = {}
#Finding neighbour with maximum occurance
for x in range(len(find_neighbours)):
response =find_neighbours[x][0]
if response in neighbour_count:
neighbour_count[response] += 1
else:
neighbour_count[response] = 1
#Select neighbour with maximum occurance in "find_neighbours" list
BestNeighbour = sorted(neighbour_count.items(), key=operator.itemgetter(1), reverse=True)
return BestNeighbour[0][0]
def FindAccuracy(test_matrix, predictions):
true_positive = 0
for i in range(len(test_matrix)):
#finding pairs of numbers which satisfies condition Predicted Number=Actual Number
if test_matrix[i][0] == predictions[i]:
true_positive += 1
#accuracy= (true_positives/total Number of test examples)*100
return (true_positive/float(len(test_matrix))) * 100.0
#Used for folding purpose
def Left_Shift_By_One(a,length_of_array):
temp=a[0]
for i in range(length_of_array-1):
a[i]=a[i+1]
a[length_of_array-1]=temp
return a
def main():
with open('mnist_train.csv', newline='') as csv_file1:
train_data_lines = csv.reader(csv_file1)
train_dataset=list(train_data_lines)
train_matrix=np.array(train_dataset).astype("int")
#prforming only on 20% of training examples
x=train_matrix[0:59999,0:785]
#10 fold cross validation
#--STEPS:
#1.shuffle training set
#2.split it into 10 arrays
#3.perform cross validation by using "Left_Shift_By_One" function
#4.At each fold calculate optimal K and its associated accuracy and append it in Final_K array and max_accuracy array.
#5.Find Final optimal k from final optimal array with max(max_accuracy)
np.random.shuffle(x)
split_matrix=np.array_split(x,10)#OR (train_matrix,10)
flag=0
Final_K=[]
Final_Accuracy=[]
for i in range(len(split_matrix)):
max_accuracy=0
optimal_K=0
#performing cross validation by using "Left_Shift_By_One" function
if(flag==0):
test_array=np.array(split_matrix[0])
train_array=split_matrix[1:10]
train_final_array=np.concatenate(train_array)
for k in range(1,11):
find_neighbours=[]
result=[]
predictions=[]
for i in range(len(test_array)):
find_neighbours=findNeighbours(train_final_array,test_array[i],k)
result = findBestNeighbour(find_neighbours)
predictions.append(result)
print('Actual Number is:' + repr(test_array[i,0])+' Predicted Number is:' + repr(result))
accuracy = FindAccuracy(test_array, predictions)
print('Accuracy: ' + repr(accuracy) + '%')
if(max_accuracy < accuracy):
max_accuracy=accuracy
optimal_K=k
else:
continue
#At each fold calculating optimal K and its associated accuracy and append it in Final_K array and max_accuracy array
Final_K.append(optimal_K)
Final_Accuracy.append(max_accuracy)
#performing cross validation by using "Left_Shift_By_One" function
if(flag==1):
test_array=[]
train_array=[]
a_array=[]
Shifted_array=Left_Shift_By_One(split_matrix,len(split_matrix))
a_array=np.array(Shifted_array).astype("int")
test_array=np.array(a_array[0])
train_array=a_array[1:10]
train_final_array=np.concatenate(train_array)
max_accuracy=0
optimal_K=0
for k in range(1,11):
find_neighbours=[]
result=[]
predictions=[]
accuracy=0
for i in range(len(test_array)):
find_neighbours=findNeighbours(train_final_array,test_array[i],k)
result = findBestNeighbour(find_neighbours)
predictions.append(result)
print('Actual Number is:' + repr(test_array[i,0])+' Predicted Number is:' + repr(result))
accuracy = FindAccuracy(test_array, predictions)
print('Accuracy: ' + repr(accuracy) + '%')
if(max_accuracy < accuracy):
max_accuracy=accuracy
optimal_K=k
else:
continue
#At each fold calculating optimal K and its associated accuracy and append it in Final_K array and max_accuracy array
Final_K.append(optimal_K)
Final_Accuracy.append(max_accuracy)
flag=1
#Find Final optimal k from final optimal array with max(max_accuracy)
Accuracy_search=np.amax(Final_Accuracy)
print("Maximum Accuracy"+repr(np.amax(Final_Accuracy)))
Index=Final_Accuracy.index(Accuracy_search)
k_final=Final_K[Index]
print("Optimal K is:"+repr(k_final))
main()