forked from scott89/mlclass
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathex8_cofi.py
163 lines (117 loc) · 4.4 KB
/
ex8_cofi.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
#!/usr/local/Cellar/python/2.7.6/bin/python
# -*- coding: utf-8 -*-
import sys
from numpy import *
import scipy.io, scipy.misc, scipy.optimize
from matplotlib import pyplot, cm, colors, lines
from mpl_toolkits.mplot3d import Axes3D
def loadMovieList():
movies = {}
counter = 0
with open('/Users/saburookita/Downloads/mlclass-ex8-004/mlclass-ex8/movie_ids.txt', 'rb') as f:
contents = f.readlines()
for content in contents:
movies[counter] = content.strip().split(' ', 1)[1]
counter += 1
return movies
def normalizeRatings( Y, R ):
m = shape( Y )[0]
Y_mean = zeros((m, 1))
Y_norm = zeros( shape( Y ) )
for i in range( 0, m ):
idx = where( R[i] == 1 )
Y_mean[i] = mean( Y[i, idx] )
Y_norm[i, idx] = Y[i, idx] - Y_mean[i]
return Y_norm, Y_mean
def unrollParams( params, num_users, num_movies, num_features ):
X = params[:num_movies * num_features]
X = X.reshape( (num_features, num_movies) ).transpose()
theta = params[num_movies * num_features:]
theta = theta.reshape( num_features, num_users ).transpose()
return X, theta
def cofiGradFunc( params, Y, R, num_users, num_movies, num_features, lamda ):
X, theta = unrollParams( params, num_users, num_movies, num_features )
inner = X.dot( theta.T ) * R - Y
X_grad = inner.dot( theta ) + lamda * X
theta_grad = inner.T.dot( X ) + lamda * theta
return r_[X_grad.T.flatten(), theta_grad.T.flatten()]
def cofiCostFunc( params, Y, R, num_users, num_movies, num_features, lamda ):
X, theta = unrollParams( params, num_users, num_movies, num_features )
J = 0.5 * sum( (X.dot( theta.T ) * R - Y) ** 2 )
regularization = 0.5 * lamda * (sum( theta**2 ) + sum(X**2))
return J + regularization
def part2_1():
mat = scipy.io.loadmat('/Users/saburookita/Downloads/mlclass-ex8-004/mlclass-ex8/ex8_movies.mat')
Y, R = mat['Y'], mat['R']
print mean( extract ( Y[0,:] * R[0,:] > 0, Y[0, :] ) )
pyplot.imshow( Y )
pyplot.ylabel( 'Movies' )
pyplot.xlabel( 'Users')
pyplot.show()
def part2_2():
mat = scipy.io.loadmat('/Users/saburookita/Downloads/mlclass-ex8-004/mlclass-ex8/ex8_movies.mat')
Y, R = mat['Y'], mat['R']
mat = scipy.io.loadmat('/Users/saburookita/Downloads/mlclass-ex8-004/mlclass-ex8/ex8_movieParams.mat')
num_features = mat['num_features']
num_users = mat['num_users']
num_movies = mat['num_movies']
X = mat['X']
theta = mat['Theta']
num_users = 4
num_features = 3
num_movies = 5
X = X[:num_movies, :num_features]
theta = theta[:num_users, :num_features]
Y = Y[:num_movies, :num_users]
R = R[:num_movies, :num_users]
params = r_[X.T.flatten(), theta.T.flatten()]
print cofiCostFunc( params, Y, R, num_users, num_movies, num_features, 0 )
print cofiGradFunc( params, Y, R, num_users, num_movies, num_features, 0 )
print cofiCostFunc( params, Y, R, num_users, num_movies, num_features, 1.5 )
print cofiGradFunc( params, Y, R, num_users, num_movies, num_features, 1.5 )
def part2_3():
movies = loadMovieList()
my_ratings = zeros((1682, 1))
my_ratings[0] = 4
my_ratings[97] = 2
my_ratings[6] = 3
my_ratings[11] = 5
my_ratings[53] = 4
my_ratings[63] = 5
my_ratings[65] = 3
my_ratings[68] = 5
my_ratings[182] = 4
my_ratings[225] = 5
my_ratings[354] = 5
# for i in range( 0, 1682 ):
# if my_ratings[i] > 0:
# print "Rated %d for %s" % (my_ratings[i], movies[i])
mat = scipy.io.loadmat('/Users/saburookita/Downloads/mlclass-ex8-004/mlclass-ex8/ex8_movies.mat')
Y, R = mat['Y'], mat['R']
Y = c_[my_ratings, Y]
R = c_[my_ratings > 0, R]
Y_norm, Y_mean = normalizeRatings( Y, R )
num_movies, num_users = shape( Y )
num_features = 10
X = random.randn( num_movies, num_features )
theta = random.randn( num_users, num_features )
initial_params = r_[X.T.flatten(), theta.T.flatten()]
lamda = 10.0
result = scipy.optimize.fmin_cg( cofiCostFunc, fprime=cofiGradFunc, x0=initial_params, \
args=( Y, R, num_users, num_movies, num_features, lamda ), \
maxiter=100, disp=True, full_output=True )
J, params = result[1], result[0]
X, theta = unrollParams( params, num_users, num_movies, num_features )
prediction = X.dot( theta.T )
my_prediction = prediction[:, 0:1] + Y_mean
idx = my_prediction.argsort(axis=0)[::-1]
my_prediction = my_prediction[idx]
for i in range(0, 10):
j = idx[i, 0]
print "Predicting rating %.1f for movie %s" % (my_prediction[j], movies[j])
def main():
# part2_1()
# part2_2()
part2_3()
if __name__ == '__main__':
main()