-
Notifications
You must be signed in to change notification settings - Fork 1
/
test_eps_score_all.m
29 lines (21 loc) · 1.39 KB
/
test_eps_score_all.m
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
% calculates the N-RMSE fit for each RBF method over a
% linear range of shape parameter values and saves the data for use later
%
% The bounds are set to be between 0.2 and 2 for all methods, because we
% are using positions that are normalized to be between 0 and 1.
%
% Copyright 2020, ETH Zurich, Multi Scale Robotics Lab, Samuel Charreyron
clear all;
load_settings;
nodes_dataset = sprintf(options.base_dataset, options.rbf_eps_grid);
% Grabbing grid number from nodes_dataset
if options.recompute ~= 0
[eps_v, scores, cond_numbers] = test_eps_score(nodes_dataset, 0, [0.1, 2], options);
save(sprintf('%s/eps_score/RBF-G-3D_%d.mat', options.data_base_path, options.rbf_eps_grid), 'eps_v', 'scores', 'cond_numbers');
[eps_v, scores, cond_numbers] = test_eps_score(nodes_dataset, 1, [0.1, 2], options);
save(sprintf('%s/eps_score/RBF-MQ-3D_%d.mat', options.data_base_path, options.rbf_eps_grid), 'eps_v', 'scores', 'cond_numbers');
[eps_v, scores, cond_numbers] = test_eps_score(nodes_dataset, 2, [0.1, 2], options);
save(sprintf('%s/eps_score/RBF-G-DF_%d.mat', options.data_base_path, options.rbf_eps_grid), 'eps_v', 'scores', 'cond_numbers');
[eps_v, scores, cond_numbers] = test_eps_score(nodes_dataset, 3, [0.1, 2], options);
save(sprintf('%s/eps_score/RBF-MQ-DF_%d.mat', options.data_base_path, options.rbf_eps_grid), 'eps_v', 'scores', 'cond_numbers');
end