-
Notifications
You must be signed in to change notification settings - Fork 2
/
expriment_realworld_data.m
287 lines (226 loc) · 9.34 KB
/
expriment_realworld_data.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
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
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
%% visulization while reading data directly from raw file
% clear;
% %Kaggle
% param.inputFile = 'Data/Kaggle/train.csv';
% param.readingNumOfTraj = 10000000; %toltalnumber 1710670
% param.img = 'Data/Portugal.png';
% param.imgMap = 'Data/Portugal.png';
% param.imgNetwork = 'Data/Portugal_network_big.png';
% if ~exist('latLong')
% [latLong,numberOfEffectiveRecords] = kaggle_trajectory_reading(param);
% end
% % param.axisRange = [-8.6986 -8.5455 41.1105 41.2574]; %boundary of Portugal(Kaggle dataset)
% param.axisRange = [-8.7024 -8.5487 41.1347 41.2459]; % for illustration
% str_save1 = 'Kaggle';
% %T-Drive
% param.inputFile = 'Data/T_Drive';
% param.readingNumOfTraj = 10000;
% param.imgMap = 'Data/beijing1.png';
% param.img = 'figs/Tdrive_raw.bmp';
% param.axisRange = [116.1935 116.5531 39.7513 40.0334 ]; % boundary of Beijing
% if ~exist('latLong')
% % [latLong,numberOfEffectiveRecords] = TDrive_trajectory_reading(param);
% load('Data/Tdrive.mat');
% end
% str_save1 = 'TDrive';
% GeoLife
param.inputFile = 'Data/GeoLife';
param.readingNumOfTraj = 100000;
% param.img = 'Data/beijing1.png';
param.img = 'figs/GeoLife_raw.bmp';
param.imgMap = 'Data/beijing1.png';
param.axisRange = [116.1935 116.5531 39.7513 40.0334 ]; % boundary of Beijing
if ~exist('latLong')
% [latLong,numberOfEffectiveRecords] = GeoLife_trajectory_reading(param);
load('Data/str_save/GeoLife_0p01_Embedding.mat');
end
% latLong = SplitOneTrajectory2Many(latLong,100);
str_save1 = 'GeoLife';
% % Hurricane
% param.inputFile = 'Data/Hurricane/Hurricane.txt';
% param.readingNumOfTraj = 10000;
% param.imgMap = 'Data/Atlantic.png';
% param.img = 'figs/Hurricane_raw.png';
% param.axisRange = [-113.7077 44.7131 5.2206 74.7249]; % boundary of Atlantic
%
% if ~exist('latLong')
% [latLong,numberOfEffectiveRecords] = Hurricane_trajectory_reading(param);
% end
% str_save1 = 'Hurricane';
% % Migration
% param.inputFile = 'Data/Migration/Swallows.csv';
% param.readingNumOfTraj = 10000;
% param.img = 'Data/SA.png';
% param.imgMap = 'Data/SA.png';
% param.axisRange = [-69.7465 -33.8018 -39.6058 16.8029]; % boundary of Atlantic
%
% if ~exist('latLong')
% [latLong,numberOfEffectiveRecords] = Migration_trajectory_reading(param);
% end
% latLong = SplitOneTrajectory2Many(latLong,3);
% str_save1 = 'Migration';
% basic clear procedure:
fprintf('number of effective records is [%d].\n original GPS coordinates number is [%d]\n',numberOfEffectiveRecords,size(latLong,1));
disp('=======specify the boundary============');
latLong(latLong(:,3) < param.axisRange(1),:) = [];
latLong(latLong(:,3) > param.axisRange(2),:) = [];
latLong(latLong(:,2) < param.axisRange(3),:) = [];
latLong(latLong(:,2) > param.axisRange(4),:) = [];
latLong(find(isnan(latLong(:,2))),:) = [];
latLong(find(isnan(latLong(:,3))),:) = [];
fprintf('number of effective records is [%d].\n original GPS coordinates number is [%d]\n',numberOfEffectiveRecords,size(latLong,1));
%%
% Specify the number of grid partition.
numberOfPartition = 2^10; % we partition the map as 1024*1024 by default,
%which means points locate in the same grid can be deemed as one simple point.
% Specify the parameters of clustering:
param.Knn = 10;
param.maxLayer = 20; % this is for manual termination.
base = 0.01;
delta = 0.01;
numOfTra = length(unique(latLong(:,1)));
num = min(20, numOfTra - 1);
iter = 2;
tensor = zeros(num,numOfTra - 1,iter);
tensor_DTW = zeros(num,numOfTra - 1,iter);
tensor_Frechet = zeros(num,numOfTra - 1,iter);
tensor_LCS = zeros(num,numOfTra - 1,iter);
tensor_EDR = zeros(num,numOfTra - 1,iter);
randnum = randperm(numOfTra);
%%
for i = 1:iter
fprintf('Start #[%d] SPs Searching\n',i);
param.epsilon_print = base + (i-1) * delta;
str_save2 = num2str(param.epsilon_print); % for the data saving usage.
idx_point = strfind(str_save2,'.');
str_save2(idx_point) = 'p';
str_save = strcat('Data/str_save/',str_save1,'_',str_save2,'.mat');
str_save_Embed = strcat('Data/str_save/',str_save1,'_',str_save2,'_Embedding_A','.mat')
try
if param.epsilon_print ~= 0.01
load(str_save_Embed);
end
catch
param.epsilon = param.epsilon_print * pi/2;
param.max_displacement = 1/numberOfPartition * pi/2; % approximate as 0.001 * pi/2
disp('=======FastSync...============');
FastSyncResult = Fast_Hierarchical_Synchronization(latLong(:,[2,3]),param);
% SPs = [ones(FastSyncResult.numberOfPrototypes,1),FastSyncResult.X];
end
[~,latLong_using_prototypes_index,SP_pairs, SP_pair_indeces] = matching_to_SPs(latLong,FastSyncResult);
[nodeCount,nodeCountIdx] = nodeCounting(SP_pairs,SP_pair_indeces);
param.visualStyle = 'heatmap';
param.visualthreshold = 1;
param.SP_pair_indeces = SP_pair_indeces;
param.nodeCount = nodeCount;
param.visualArrow = true;
param.isNodeColor = false;
visualization(SP_pairs,param);
%% DTW
fprintf('Start #[%d] round DTW calculation\n',i);
[~, TopK_record_DTW] = Tra_distance_matrix_using_dtw_index_randomK(latLong_using_prototypes_index, FastSyncResult.X,randnum,num);
disp('DTW : 1. distance matrix done!');
tensor_DTW(:,:,i) = TopK_record_DTW;
%% LCS
fprintf('Start #[%d] round LCS calculation\n',i);
[~, TopK_record_LCS] = Tra_distance_matrix_using_LCS_index_randomK(latLong_using_prototypes_index, FastSyncResult.X,randnum,num);
disp('DTW : 1. distance matrix done!');
tensor_LCS(:,:,i) = TopK_record_LCS;
%% Frechet distance
fprintf('Start #[%d] round Frechet calculation\n',i);
[~, TopK_record_Frechet] = Tra_distance_matrix_using_Frechet_index_randomK(latLong_using_prototypes_index, FastSyncResult.X,randnum,num);
disp('DTW : 1. distance matrix done!');
tensor_Frechet(:,:,i) = TopK_record_Frechet;
%% EDR
fprintf('Start #[%d] round EDR calculation\n',i);
tol = max(param.axisRange(2) - param.axisRange(1), param.axisRange(4) - param.axisRange(3))/10;
[~, TopK_record_EDR] = Tra_distance_matrix_using_EDR_index_randomK(latLong_using_prototypes_index, FastSyncResult.X,randnum,num,tol);
disp('DTW : 1. distance matrix done!');
tensor_EDR(:,:,i) = TopK_record_EDR;
%% SP Embedding
fprintf('Start #[%d] embedding\n',i);
param.dim = 50;
param.IterTime = 1;
param.NegativeSamples = 3;
param.learningRate = 0.025;
param.window = 2;
param.similarNum = 0;
param.bin = 5;
param.degreeRate = 0.8;
[S, T, ~] = AsymmetricSPEmbedding(SP_pair_indeces, latLong_using_prototypes_index, param);
SPEmbeddedVector = [S,T];
disp('Embedding done!');
% Trajectory embedding using sum
[TrajEmbeddedVector,index_of_trajectory] = TrajectoryEmbedding(SPEmbeddedVector, latLong_using_prototypes_index,'sum');
fprintf('Start #[%d] round neighbors searching\n',i);
%% Similarity Using Embedding
TopK_record = zeros(num,numOfTra - 1);
for j = 1:num
idx = randnum(j);
% Euclidean cosine
[~,TopK_index] = findSimilarVectors(TrajEmbeddedVector,numOfTra - 1,idx,'Euclidean');
TopK_record(j,:) = TopK_index';
end
tensor(:,:,i) = TopK_record;
% %% Calculate using second order
%
% latLong2SP = [latLong(:,1),FastSyncResult.Map_Original_To_Prototype];
% [~,maxIdx]= findSimilarTra_secondOrder(FastSyncResult.X,latLong_using_prototypes_index,latLong2SP,num,numOfTra - 1,randnum, 'binary');
% tensor(:,:,i) = maxIdx;
%
%
save(str_save_Embed);
disp('save successfully!');
end
%%
fprintf('Start calculate consistency\n');
%% Embedding
intersetNum = zeros(num, numOfTra-1);
for k = 1:num
mat = reshape(tensor(k,:,:),numOfTra-1,iter)';
numberVec = countIntersect(mat);
intersetNum(k,:) = numberVec;
end
drawData = mean(intersetNum)./(1:numOfTra-1);
%% DTW
intersetNum_DTW = zeros(num, numOfTra-1);
for k = 1:num
mat = reshape(tensor_DTW(k,:,:),numOfTra-1,iter)';
numberVec = countIntersect(mat);
intersetNum_DTW(k,:) = numberVec;
end
drawData_DTW = mean(intersetNum_DTW)./(1:numOfTra-1);
%% LCS
intersetNum_LCS = zeros(num, numOfTra-1);
for k = 1:num
mat = reshape(tensor_LCS(k,:,:),numOfTra-1,iter)';
numberVec = countIntersect(mat);
intersetNum_LCS(k,:) = numberVec;
end
drawData_LCS = mean(intersetNum_LCS)./(1:numOfTra-1);
%% Frechet
intersetNum_Frechet = zeros(num, numOfTra-1);
for k = 1:num
mat = reshape(tensor_Frechet(k,:,:),numOfTra-1,iter)';
numberVec = countIntersect(mat);
intersetNum_Frechet(k,:) = numberVec;
end
drawData_Frechet = mean(intersetNum_Frechet)./(1:numOfTra-1);
%% EDR
intersetNum_EDR = zeros(num, numOfTra-1);
for k = 1:num
mat = reshape(tensor_EDR(k,:,:),numOfTra-1,iter)';
numberVec = countIntersect(mat);
intersetNum_EDR(k,:) = numberVec;
end
drawData_EDR = mean(intersetNum_EDR)./(1:numOfTra-1);
fprintf('Done!\n');
figure;
hold on;
plot(1:numOfTra-1, drawData);
plot(1:numOfTra-1, drawData_DTW);
plot(1:numOfTra-1, drawData_Frechet);
plot(1:numOfTra-1, drawData_LCS);
plot(1:numOfTra-1, drawData_EDR);
hold off;
legend('Embedding','DTW','Frechet','LCS','EDR');