forked from ofekfeller/FinalSoftwareProject
-
Notifications
You must be signed in to change notification settings - Fork 0
/
spkmeans.py
81 lines (54 loc) · 1.93 KB
/
spkmeans.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
import sys
import numpy as np
import pandas as pd
import spkmeansmodule
def read_csv_files_to_numpy(filename1):
file = pd.read_csv(filename1, header=None)
return file.to_numpy()
def find_min_d(vec, centers):
min_val = np.infty
for i in range(len(centers)):
norm = np.linalg.norm(np.subtract(centers[i], vec))
norm = norm**2
if norm < min_val:
min_val = norm
return min_val
# is there better approach we can handle this with?
def is_int(num):
int(num)
def get_args(args):
if len(args) == 4:
return int(args[1]), args[2], args[3]
else:
raise Exception("Arguments are corrupted")
def main():
np.random.seed(0)
_K, goal, filename = get_args(sys.argv)
points = read_csv_files_to_numpy(filename)
if goal != "spk":
spkmeansmodule.execute_goal(points.tolist(), goal)
return
assert (_K < points.shape[0])
spk_points = spkmeansmodule.spk_points(points.tolist(), [1 for i in range(_K)])
spk_points = np.asarray(spk_points)
_K = spk_points.shape[1] # the new K is the number of columns
z = 1
points_num = spk_points.shape[0]
rnd = np.random.choice(range(points_num))
indexes = str(rnd)
centers = [spk_points[rnd, :]]
for i in range(_K-1):
prob_vec = [find_min_d(spk_points[j, :], centers) for j in range(points_num)]
total = sum(prob_vec)
prob_vec = [val/total for val in prob_vec]
p = np.random.choice([i for i in range(points_num)], p=prob_vec)
indexes += f",{p}"
centers.append(spk_points[p, :])
print(indexes)
centers = [center.tolist() for center in centers]
spkmeansmodule.fit(spk_points.tolist(), centers)
#res = np.round(np.array(res), decimals=4)
#for row in res:
# print(",".join(map(str, row)))
if __name__ == '__main__':
main()