-
Notifications
You must be signed in to change notification settings - Fork 0
/
mccaffrey_neural.py
341 lines (286 loc) · 11.6 KB
/
mccaffrey_neural.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
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
"""
from https://visualstudiomagazine.com/articles/2014/12/01/back-propagation-using-python.aspx
"""
# neuralnetbackprop.py
# uses Python version 2.7.8
import random
import math
# ------------------------------------
def show_data(matrix, num_first_rows):
#for i in range(len(matrix)):
for i in range(0, num_first_rows-1):
print "[" + str(i).rjust(2) + "]",
for j in range(len(matrix[i])):
print str("%.1f" % matrix[i][j]).rjust(5),
print "\n",
print "........"
last_row = len(matrix) - 1
print "[" + str(last_row).rjust(2) + "]",
for j in range(len(matrix[last_row])):
print str("%.1f" % matrix[last_row][j]).rjust(5),
print "\n"
def show_vector(vector):
for i in range(len(vector)):
if i % 8 == 0: # 8 columns
print "\n",
if vector[i] >= 0.0:
print '',
print "%.4f" % vector[i], # 4 decimals
print "\n"
# ------------------------------------
class NeuralNetwork:
def __init__(self, num_input, num_hidden, num_output):
self.num_input = num_input
self.num_hidden = num_hidden
self.num_output = num_output
self.inputs = [0 for i in range(num_input)]
self.ih_weights = self.make_matrix(num_input, num_hidden)
self.h_biases = [0 for i in range(num_hidden)]
self.h_outputs = [0 for i in range(num_hidden)]
self.ho_weights = self.make_matrix(num_hidden, num_output)
self.o_biases = [0 for i in range(num_output)]
self.outputs = [0 for i in range(num_output)]
# random.seed(0) # hidden function is 'normal' approach
self.rnd = random.Random(0) # allows multiple instances
self.initialize_weights()
def make_matrix(self, rows, cols):
result = [[0 for j in range(cols)] for i in range(rows)]
return result
def set_weights(self, weights):
k = 0
for i in range(self.num_input):
for j in range(self.num_hidden):
self.ih_weights[i][j] = weights[k]
k += 1
for i in range(self.num_hidden):
self.h_biases[i] = weights[k]
k += 1
for i in range(self.num_hidden):
for j in range(self.num_output):
self.ho_weights[i][j] = weights[k]
k += 1
for i in range(self.num_output):
self.o_biases[i] = weights[k]
k += 1
def get_weights(self):
num_wts = ((self.num_input * self.num_hidden) + self.num_hidden +
(self.num_hidden * self.num_output) + self.num_output)
result = [0 for i in range(num_wts)]
k = 0
for i in range(self.num_input):
for j in range(self.num_hidden):
result[k] = self.ih_weights[i][j]
k += 1
for i in range(self.num_hidden):
result[k] = self.h_biases[i]
k += 1
for i in range(self.num_hidden):
for j in range(self.num_output):
result[k] = self.ho_weights[i][j]
k += 1
for i in range(self.num_output):
result[k] = self.o_biases[i]
k += 1
return result
def initialize_weights(self):
num_wts = ((self.num_input * self.num_hidden) + self.num_hidden +
(self.num_hidden * self.num_output) + self.num_output)
wts = [0 for i in range(num_wts)]
lo = -0.01
hi = 0.01
for i in range(len(wts)):
wts[i] = (hi - lo) * self.rnd.random() + lo
self.set_weights(wts)
def compute_outputs(self, x_values):
h_sums = [0 for i in range(self.num_hidden)]
o_sums = [0 for i in range(self.num_output)]
for i in range(len(x_values)):
self.inputs[i] = x_values[i]
for j in range(self.num_hidden):
for i in range(self.num_input):
h_sums[j] += (self.inputs[i] * self.ih_weights[i][j])
for i in range(self.num_hidden):
h_sums[i] += self.h_biases[i]
for i in range(self.num_hidden):
self.h_outputs[i] = self.hypertan(h_sums[i])
for j in range(self.num_output):
for i in range(self.num_hidden):
o_sums[j] += (self.h_outputs[i] * self.ho_weights[i][j])
for i in range(self.num_output):
o_sums[i] += self.o_biases[i]
soft_out = self.softmax(o_sums)
for i in range(self.num_output):
self.outputs[i] = soft_out[i]
result = [0 for i in range(self.num_output)]
for i in range(self.num_output):
result[i] = self.outputs[i]
return result
def hypertan(self, x):
if x < -20.0:
return -1.0
elif x > 20.0:
return 1.0
else:
return math.tanh(x)
def softmaxnaive(self, o_sums):
div = 0
for i in range(len(o_sums)):
div = div + math.exp(o_sums[i])
result = [0 for i in range(len(o_sums))]
for i in range(len(o_sums)):
result[i] = math.exp(o_sums[i]) / div
return result
def softmax(self, o_sums):
m = max(o_sums)
scale = 0
for i in range(len(o_sums)):
scale = scale + (math.exp(o_sums[i] - m))
result = [0 for i in range(len(o_sums))]
for i in range(len(o_sums)):
result[i] = math.exp(o_sums[i] - m) / scale
return result
def train(self, train_data, max_epochs, learn_rate, momentum):
o_grads = [0 for i in range(self.num_output)] # gradients
h_grads = [0 for i in range(self.num_hidden)]
ih_prev_weights_delta = self.make_matrix(num_input, num_hidden) # momentum
h_prev_biases_delta = [0 for i in range(self.num_hidden)]
ho_prev_weights_delta = self.make_matrix(num_hidden, num_output)
o_prev_biases_delta = [0 for i in range(self.num_output)]
epoch = 0
x_values = [0 for i in range(self.num_input)]
t_values = [0 for i in range(self.num_output)]
sequence = [i for i in range(len(train_data))]
while epoch < max_epochs:
self.rnd.shuffle(sequence)
for ii in range(len(train_data)):
idx = sequence[ii]
for j in range(self.num_input): # peel off x_values
x_values[j] = train_data[idx][j]
for j in range(self.num_output): # peel off t_values
t_values[j] = train_data[idx][j + self.num_input]
self.compute_outputs(x_values) # outputs stored internally
# --- update-weights (back-prop) section
for i in range(self.num_output): # 1. compute output gradients
derivative = (1 - self.outputs[i]) * self.outputs[i]
o_grads[i] = derivative * (t_values[i] - self.outputs[i])
for i in range(self.num_hidden): # 2. compute hidden gradients
derivative = (1 - self.h_outputs[i]) * (1 + self.h_outputs[i])
sum = 0
for j in range(self.num_output):
x = o_grads[j] * self.ho_weights[i][j]
sum += x
h_grads[i] = derivative * sum
for i in range(self.num_input): # 3a. update input-hidden weights
for j in range(self.num_hidden):
delta = learn_rate * h_grads[j] * self.inputs[i]
self.ih_weights[i][j] += delta
self.ih_weights[i][j] += momentum * ih_prev_weights_delta[i][j] # momentum
ih_prev_weights_delta[i][j] = delta # save the delta for momentum
for i in range(self.num_hidden): # 3b. update hidden biases
delta = learn_rate * h_grads[i]
self.h_biases[i] += delta
self.h_biases[i] += momentum * h_prev_biases_delta[i]; # momentum
h_prev_biases_delta[i] = delta # save the delta
for i in range(self.num_hidden): # 4a. update hidden-output weights
for j in range(self.num_output):
delta = learn_rate * o_grads[j] * self.h_outputs[i]
self.ho_weights[i][j] += delta
self.ho_weights[i][j] += momentum * ho_prev_weights_delta[i][j]; # momentum
ho_prev_weights_delta[i][j] = delta # save
for i in range(self.num_output): # 4b. update output biases
delta = learn_rate * o_grads[i]
self.o_biases[i] += delta
self.o_biases[i] += momentum * o_prev_biases_delta[i] # momentum
o_prev_biases_delta[i] = delta # save
# --- end update-weights
epoch += 1
result = self.get_weights()
return result
def accuracy(self, data):
num_correct = 0
num_wrong = 0
x_values = [0 for i in range(self.num_input)]
t_values = [0 for i in range(self.num_output)]
for i in range(len(data)):
for j in range(self.num_input): # peel off x_values
x_values[j] = data[i][j]
for j in range(self.num_output): # peel off t_values
t_values[j] = data[i][j + self.num_input]
y_values = self.compute_outputs(x_values)
max_index = y_values.index(max(y_values))
if t_values[max_index] == 1.0:
num_correct += 1;
else:
num_wrong += 1;
return (num_correct * 1.0) / (num_correct + num_wrong)
# ------------------------------------
print "\nBegin neural network using Python demo"
print "\nGoal is to predict species from color, petal length, petal width \n"
print "The 30-item raw data looks like: \n"
print "[0] blue, 1.4, 0.3, setosa"
print "[1] pink, 4.9, 1.5, versicolor"
print "[2] teal, 5.6, 1.8, virginica"
print ". . ."
print "[29] pink, 5.9, 1.5, virginica"
train_data = ([[0 for j in range(7)]
for i in range(24)]) # 24 rows, 7 cols
train_data[0] = [ 1, 0, 1.4, 0.3, 1, 0, 0 ]
train_data[1] = [ 0, 1, 4.9, 1.5, 0, 1, 0 ]
train_data[2] = [ -1, -1, 5.6, 1.8, 0, 0, 1 ]
train_data[3] = [ -1, -1, 6.1, 2.5, 0, 0, 1 ]
train_data[4] = [ 1, 0, 1.3, 0.2, 1, 0, 0 ]
train_data[5] = [ 0, 1, 1.4, 0.2, 1, 0, 0 ]
train_data[6] = [ 1, 0, 6.6, 2.1, 0, 0, 1 ]
train_data[7] = [ 0, 1, 3.3, 1.0, 0, 1, 0 ]
train_data[8] = [ -1, -1, 1.7, 0.4, 1, 0, 0 ]
train_data[9] = [ 0, 1, 1.5, 0.1, 0, 1, 1 ]
train_data[10] = [ 0, 1, 1.4, 0.2, 1, 0, 0 ]
train_data[11] = [ 0, 1, 4.5, 1.5, 0, 1, 0 ]
train_data[12] = [ 1, 0, 1.4, 0.2, 1, 0, 0 ]
train_data[13] = [ -1, -1, 5.1, 1.9, 0, 0, 1 ]
train_data[14] = [ 1, 0, 6.0, 2.5, 0, 0, 1 ]
train_data[15] = [ 1, 0, 3.9, 1.4, 0, 1, 0 ]
train_data[16] = [ 0, 1, 4.7, 1.4, 0, 1, 0 ]
train_data[17] = [ -1, -1, 4.6, 1.5, 0, 1, 0 ]
train_data[18] = [ -1, -1, 4.5, 1.7, 0, 0, 1 ]
train_data[19] = [ 0, 1, 4.5, 1.3, 0, 1, 0 ]
train_data[20] = [ 1, 0, 1.5, 0.2, 1, 0, 0 ]
train_data[21] = [ 0, 1, 5.8, 2.2, 0, 0, 1 ]
train_data[22] = [ 0, 1, 4.0, 1.3, 0, 1, 0 ]
train_data[23] = [ -1, -1, 5.8, 1.8, 0, 0, 1 ]
test_data = ([[0 for j in range(7)]
for i in range(6)]) # 6 rows, 7 cols
test_data[0] = [ 1, 0, 1.5, 0.2, 1, 0, 0 ]
test_data[1] = [ -1, -1, 5.9, 2.1, 0, 0, 1 ]
test_data[2] = [ 0, 1, 1.4, 0.2, 1, 0, 0 ]
test_data[3] = [ 0, 1, 4.7, 1.6, 0, 1, 0 ]
test_data[4] = [ 1, 0, 4.6, 1.3, 0, 1, 0 ]
test_data[5] = [ 1, 0, 6.3, 1.8, 0, 0, 1 ]
print "\nFirst few lines of encoded training data are: \n"
show_data(train_data, 4)
print "\nThe encoded test data is: \n"
show_data(test_data, 5)
print "\nCreating a 4-input, 5-hidden, 3-output neural network"
print "Using tanh and softmax activations \n"
num_input = 4
num_hidden = 5
num_output = 3
nn = NeuralNetwork(num_input, num_hidden, num_output)
max_epochs = 70 # artificially small
learn_rate = 0.08 # artificially large
momentum = 0.01
print "Setting max_epochs = " + str(max_epochs)
print "Setting learn_rate = " + str(learn_rate)
print "Setting momentum = " + str(momentum)
print "\nBeginning training using back-propagation"
weights = nn.train(train_data, max_epochs, learn_rate, momentum)
print "Training complete \n"
print "Final neural network weights and bias values:"
show_vector(weights)
print "Model accuracy on training data =",
acc_train = nn.accuracy(train_data)
print "%.4f" % acc_train
print "Model accuracy on test data =",
acc_test = nn.accuracy(test_data)
print "%.4f" % acc_test
print "\nEnd back-prop demo \n"