forked from M-Lin-DM/Uniform-Orientation-Sampling
-
Notifications
You must be signed in to change notification settings - Fork 0
/
plots.py
31 lines (28 loc) · 1.4 KB
/
plots.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
import numpy as np
from Uniform_Orientation_Sampling import UniformOrientationSampling
from Plotting_tools import NDScatter
from Evaluations import KNNDistanceDistribution
Pops = [int(j) for j in np.logspace(1, 5, num=5, endpoint=True, base=5)]
dim_range = np.arange(2, 6, 1)
save_path_emb = 'C:/Users/mirbe/Documents/Experiments/Uniform-Orientation-Sampling/figures/emb3D/'
save_path_naive = 'C:/Users/mirbe/Documents/Experiments/Uniform-Orientation-Sampling/figures/emb3Dnaive/'
save_path_hist = 'C:/Users/mirbe/Documents/Experiments/Uniform-Orientation-Sampling/figures/histograms/'
# for s in Pops:
#
# UOS = UniformOrientationSampling(dimensions=3, pop_size=s, iterations=1500, approach='KNN_repulsion')
# embedding = UOS.run_optimizer()
#
# # for plotting optimized embedding
# plotter = NDScatter(embedding, UOS, make_raster=True, save_path=save_path_emb)
# plotter.scatter3D_mpl(plot_cap=True)
#
# KNNDD = KNNDistanceDistribution(embedding, UOS, save_path_naive=save_path_naive, save_path_histogram=save_path_hist)
# KNNDD.get_KNNDD()
#
# KNNDD.save_naive_emb()
# ----------------------
for d in dim_range:
UOS = UniformOrientationSampling(dimensions=d, pop_size=400, iterations=1500, approach='KNN_repulsion')
embedding = UOS.run_optimizer()
KNNDD = KNNDistanceDistribution(embedding, UOS, save_path_histogram=save_path_hist)
KNNDD.get_KNNDD()