UniversityMachineLearningAssignment1/data_gen.py at master · AdamZA/UniversityMachineLearningAssignment1 · GitHub

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
#!/usr/bin/python
'''
Created on 14 Feb 2014

@author: benjamin
'''
import numpy as np

no_training_samples_per_dim = 50
no_inputs = 2
no_hidden = 150
no_out = 2
learning_rate = 0.075
mse_cutoff = 0.1
file_name = "training_data.txt"


def normalize(v):
    return v / v.max()
'''
generate many possible responses for the dot product
between the look direction and the direction to the mine/
super mine/rock.

Remember (first year / CS Games) the dot product between
two normalized vectors ranges from -1 through to 1:
+1 if the two vectors point in the same direction
0 if the two vectors are orthogonal
-1 if the two vectors point in opposite directions
'''
dot_look_mine_vec = np.linspace(-1,1,no_training_samples_per_dim)
dot_look_super_mine_or_rock_vec = np.linspace(-1,1,no_training_samples_per_dim)
#if we're pointing towards the mine do not turn:
resp_dot_look_mine_vec = np.ones(len(dot_look_mine_vec)) #always try to turn closer to the direction of the mine
#for i in range(0,len(dot_look_mine_vec)):
#    resp_dot_look_mine_vec[i] = 1 if dot_look_mine_vec[i] < 0.95 else 0
#if we're pointing towards the super mine / rock turn sharply,
#if we're pointing somewhere between perpendicular and the same direction stop turning:
resp_dot_look_super_mine_or_rock_vec = np.zeros(len(dot_look_super_mine_or_rock_vec))
for i in range(0,len(dot_look_super_mine_or_rock_vec)):
    resp_dot_look_super_mine_or_rock_vec[i] = 1 if dot_look_super_mine_or_rock_vec[i] > 0.45 else 0

f = open(file_name,'w')
f.write(str(no_training_samples_per_dim**2)+"\n")
f.write(str(no_inputs)+"\n")
f.write(str(no_hidden)+"\n")
f.write(str(no_out)+"\n")
f.write(str(learning_rate)+"\n")
f.write(str(mse_cutoff)+"\n")
for m in range(0,no_training_samples_per_dim):
    print "Dumping state space ... %f%% complete" % (m/float(no_training_samples_per_dim)*100.0)
    #for r in range(0,no_training_samples_per_dim):
    for sm in range(0,no_training_samples_per_dim):
        f.write(str(dot_look_mine_vec[m]) + " " +
                str(dot_look_super_mine_or_rock_vec[sm]) + "\n")
        f.write(str(resp_dot_look_mine_vec[m] if resp_dot_look_super_mine_or_rock_vec[sm] == 0 else 0) + " " +
                str(resp_dot_look_super_mine_or_rock_vec[sm]) + "\n")
print "Complete... Terminating"
f.close()