-
Notifications
You must be signed in to change notification settings - Fork 0
/
randomLHSGeneratorDrift.py
50 lines (37 loc) · 2.15 KB
/
randomLHSGeneratorDrift.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
__author__ = 'Daniel Puschmann'
from randomLHSGenerator import RandomLHSGenerator
from random import Random
class RandomLHSGeneratorDrift(RandomLHSGenerator):
def __init__(self, speed=1, random_seed_generation=1, dimensions=10, center_num=10,
sample_limit=None, class_num=2,
hyper_cube_length=500, variance=60, file_name = None,
stream_name='LHSData', criterion='maximin', distribution='gauss'):
super(RandomLHSGeneratorDrift, self).__init__(random_seed_generation, dimensions,
center_num, sample_limit, class_num,
hyper_cube_length, variance, file_name,
stream_name, criterion, distribution)
self.speed = speed
self.generateCentroids()
self.rand_model = Random()
self.drift_times = [self.rand_model.randint(50, 2000) for i in xrange(dimensions)]
def change_distribution(self, distribution):
self.feature_distributions = distribution
def shift_centroid(self):
for centre_num, centre in enumerate(self.centroids.centres):
for dimension, value in enumerate(centre):
if self.iteration > self.drift_times[dimension]:
if self.rand_model.randint(0, 1) is 1:
#shift centre of feature
self.centroids.centres[centre_num] = self.rand_model.random() * self.hyper_cube_length
if self.rand_model.randint(0, 1) is 1:
#shift distribution
self.feature_distributions[dimension] = self.supported_distributions[self.rand_model.randrange(0, self.number_of_distributions)]
if __name__ == '__main__':
features = [2,3,4,5]
file_names = ["LHSGeneratedK100F%i" % f for f in features]
limit = 100000
for f, fname in zip(features, file_names):
stream = RandomLHSGeneratorDrift(file_name=fname, sample_limit=limit, center_num=100, dimensions=f)
for i, instance in enumerate(stream):
if i % 10000 == 0:
print "Instance: %i" % i