-
Notifications
You must be signed in to change notification settings - Fork 0
/
RepresentativekNN.m
389 lines (383 loc) · 15.1 KB
/
RepresentativekNN.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
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
% Script to apply repknn algorithm on real data of sound-based localization
% input:
% 272 set of 10000 sound measurements at a sample rate of 10 kHz (1 s)
% - 80 training sets (5 for each label from 1 to 16, datasets 1-80)
% - 192 test sets for prediction (12 for each area, datasets 81-272)
% - 272 coordinates, 1 for each data set, sorted by time of measurement
% output:
% - plot of label-depending colored training datasets (squares) and
% predicted test datasets (circles) with number of label
% parameter:
% - useToolbox: 1 to use Matlab Toolbox 0 for COSY kNN & kMeans
% - representatives: set number of representatives (1<representatives<5)
% - neighbors: set number of neighbors (>0)
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% clear workspace
clear all;
useToolbox = 0;
representatives = 3;
neighbors = 1;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% GET DATA
% load coordinates
Coord = load('koords_paper.csv');
% path to files with datasets - please check
path = 'dataPaper\';
% get all fileproperties in directory
files = dir(path);
% get filenames
filenames = {files.name};
% count files
amountFiles = (numel(files)-2);
% declare help variables length raw data, used data
lengthRawData = 10000;
lengthData = 2000;
% declare data containing matrices
RawData = zeros(lengthRawData, amountFiles);
Data = zeros(lengthData, amountFiles);
% move raw data from files to raw data matrix
% for each file/column
for k = 1 : amountFiles
% create path from prefix and filename
fullpath = strcat(path,filenames{k+2});
% load raw data
RawData(:,k) = load(fullpath);
end
% sensor sends data with offset, get mean from each raw dataset to adjust
means = mean(RawData);
% adjust each raw dataset by mean value
for i = 1:amountFiles
RawData(:,i) = RawData(:,i) - means(:,i);
end
% some datasets containing noise at beginning, clear these parts
RawData([1:400],5) = 0;
RawData([1:400],26) = 0;
RawData([1:400],34) = 0;
RawData([1:400],63) = 0;
RawData([1:400],182) = 0;
RawData([1:400],224) = 0;
% filter timeframe of 2000 samples (200 ms) from raw data for each file
for j = 1:amountFiles
% check values for threshold
for i = 1:lengthRawData
% if absolut value bigger 50, save index
if (abs(RawData(i,j)) > 50)
z = i - 1;
break;
end
end
% move raw data from threshold
for h = 1:lengthData
Data(h,j) = RawData((z+h-3),j);
end
% normalize data
Data(:,j) = Data(:,j)/norm(Data(:,j));
z = 0;
end
% END GET DATA
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% SET TRAINING DATA
% set size training data
train = 80;
% number of training clusters
parts = 16;
% units per cluster
trainunits = train/parts;
% create labelarray, 1-16 for each training area, 0 for testsets
label = zeros(1,amountFiles);
for i = 1 : parts
for j = 1 : trainunits
label(((i-1)*trainunits)+j) = i;
end
end
% create trainingdatasets
M1 = [];
M2 = [];
M3 = [];
M4 = [];
M5 = [];
M6 = [];
M7 = [];
M8 = [];
M9 = [];
M10 = [];
M11 = [];
M12 = [];
M13 = [];
M14 = [];
M15 = [];
M16 = [];
% sort training data depending on label
for i = 1:train
if (label(i)==1)
M1 = [M1, Data(:,i)];
elseif (label(i)==2)
M2 = [M2, Data(:,i)];
elseif (label(i)==3)
M3 = [M3, Data(:,i)];
elseif (label(i)==4)
M4 = [M4, Data(:,i)];
elseif (label(i)==5)
M5 = [M5, Data(:,i)];
elseif (label(i)==6)
M6 = [M6, Data(:,i)];
elseif (label(i)==7)
M7 = [M7, Data(:,i)];
elseif (label(i)==8)
M8 = [M8, Data(:,i)];
elseif (label(i)==9)
M9 = [M9, Data(:,i)];
elseif (label(i)==10)
M10 = [M10, Data(:,i)];
elseif (label(i)==11)
M11 = [M11, Data(:,i)];
elseif (label(i)==12)
M12 = [M12, Data(:,i)];
elseif (label(i)==13)
M13 = [M13, Data(:,i)];
elseif (label(i)==14)
M14 = [M14, Data(:,i)];
elseif (label(i)==15)
M15 = [M15, Data(:,i)];
elseif (label(i)==16)
M16 = [M16, Data(:,i)];
end
end
% END SET TRAINING DATA
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% GET REPRESENTATIVES FROM TRAINING DATA
% clustering to get representatives for each label
if useToolbox == 1
[idx1,R1] = kmeans(M1',representatives,'Distance','cosine');
[idx2,R2] = kmeans(M2',representatives,'Distance','cosine');
[idx3,R3] = kmeans(M3',representatives,'Distance','cosine');
[idx4,R4] = kmeans(M4',representatives,'Distance','cosine');
[idx5,R5] = kmeans(M5',representatives,'Distance','cosine');
[idx6,R6] = kmeans(M6',representatives,'Distance','cosine');
[idx7,R7] = kmeans(M7',representatives,'Distance','cosine');
[idx8,R8] = kmeans(M8',representatives,'Distance','cosine');
[idx9,R9] = kmeans(M9',representatives,'Distance','cosine');
[idx10,R10] = kmeans(M10',representatives,'Distance','cosine');
[idx11,R11] = kmeans(M11',representatives,'Distance','cosine');
[idx12,R12] = kmeans(M12',representatives,'Distance','cosine');
[idx13,R13] = kmeans(M13',representatives,'Distance','cosine');
[idx14,R14] = kmeans(M14',representatives,'Distance','cosine');
[idx15,R15] = kmeans(M15',representatives,'Distance','cosine');
[idx16,R16] = kmeans(M16',representatives,'Distance','cosine');
else
% clustering to get representatives for each label
[R1] = kMeansCOSY(M1', representatives, 1);
[R2] = kMeansCOSY(M2', representatives, 1);
[R3] = kMeansCOSY(M3', representatives, 1);
[R4] = kMeansCOSY(M4', representatives, 1);
[R5] = kMeansCOSY(M5', representatives, 1);
[R6] = kMeansCOSY(M6', representatives, 1);
[R7] = kMeansCOSY(M7', representatives, 1);
[R8] = kMeansCOSY(M8', representatives, 1);
[R9] = kMeansCOSY(M9', representatives, 1);
[R10] = kMeansCOSY(M10', representatives, 1);
[R11] = kMeansCOSY(M11', representatives, 1);
[R12] = kMeansCOSY(M12', representatives, 1);
[R13] = kMeansCOSY(M13', representatives, 1);
[R14] = kMeansCOSY(M14', representatives, 1);
[R15] = kMeansCOSY(M15', representatives, 1);
[R16] = kMeansCOSY(M16', representatives, 1);
end
% Repräsentanten normieren
for j = 1 : representatives
R1(j,:) = R1(j,:)/norm(R1(j,:));
R2(j,:) = R2(j,:)/norm(R2(j,:));
R3(j,:) = R3(j,:)/norm(R3(j,:));
R4(j,:) = R4(j,:)/norm(R4(j,:));
R5(j,:) = R5(j,:)/norm(R5(j,:));
R6(j,:) = R6(j,:)/norm(R6(j,:));
R7(j,:) = R7(j,:)/norm(R7(j,:));
R8(j,:) = R8(j,:)/norm(R8(j,:));
R9(j,:) = R9(j,:)/norm(R9(j,:));
R10(j,:) = R10(j,:)/norm(R10(j,:));
R11(j,:) = R11(j,:)/norm(R11(j,:));
R12(j,:) = R12(j,:)/norm(R12(j,:));
R13(j,:) = R13(j,:)/norm(R13(j,:));
R14(j,:) = R14(j,:)/norm(R14(j,:));
R15(j,:) = R15(j,:)/norm(R15(j,:));
R16(j,:) = R16(j,:)/norm(R16(j,:));
end
% END GET REPRESENTATIVES FROM TRAINING DATA
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% PREDICT TEST DATA
% set label for representatives
sizelabel2 = representatives*parts;
label2 = ones(1,sizelabel2);
for i = 1 : parts
for j = 1 : representatives
label2(((i-1)*representatives)+j) = i; % 6-10
end
end
if useToolbox == 1
% apply k-nearest neighbors with Toolbox
mdl = fitcknn([R1;R2;R3;R4;R5;R6;R7;R8;R9;R10;R11;R12;R13;R14;R15;R16],label2,'Distance','cosine');
% set number auf neighbors
mdl.NumNeighbors = neighbors;
mdl.Distance = 'cosine';
mdl.BreakTies = 'nearest';
dist = mdl.Distance;
else
distance = 1;
if distance == 1
dist = 'cosine';
elseif distance == 2
dist = 'euclidean';
elseif distance == 3
dist = 'sqeuclidean';
end
end
% END PREDICT TEST DATA
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
quote = 0;
tested = 0;
correct = 0;
s = zeros(1, length(Data(:,1)));
for y = 1:amountFiles
if (label(y) == 0)
tested = tested + 1;
if useToolbox == 1
% set s to the label of the next representatives to actual dataset
s(y) = predict(mdl,Data(:,y)');
else
R = [R1;R2;R3;R4;R5;R6;R7;R8;R9;R10;R11;R12;R13;R14;R15;R16];
s(y) = kNNCOSY(R,label2,neighbors,Data(:,y),1);
end
areal = fix((y-81)/12) + 1;
if(s(y) == areal)
correct = correct + 1;
end
end
end
quote=correct/tested;
% PLOT RESULTS
figure('Name',dist);
if useToolbox == 1
title(sprintf('Representative kNN with Toolbox\nCorrect: %.2f', quote*100));
else
title(sprintf('Representative kNN without Toolbox\nCorrect: %.2f', quote*100));
end
hold
% set size and position of figure
set(gcf, 'Position', [0, 0, 1000, 1000])
xlabel('[cm]');
ylabel('[cm]');
% plot grid
set(gca,'xtick',[0:15:60])
set(gca,'ytick',[0:15:60])
grid on
% offsets to adjust numbers to circles and squares
xoffset = 0.52;
yoffset = 0.1;
% check all files
for y = 1:amountFiles
% if no training data
if (label(y) == 0)
% plot color and number at position of dataset depending on s
if (s(y)==1)
plot(Coord(y,1),Coord(y,2),'o', 'LineWidth', 2, 'MarkerSize', 18, 'Color', 'r');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'13')
elseif (s(y)==2)
plot(Coord(y,1),Coord(y,2),'o', 'LineWidth', 2, 'MarkerSize', 18, 'Color', 'b');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'9')
elseif (s(y)==3)
plot(Coord(y,1),Coord(y,2),'o', 'LineWidth', 2, 'MarkerSize', 18, 'Color', 'g');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'5')
elseif (s(y)==4)
plot(Coord(y,1),Coord(y,2),'o', 'LineWidth', 2, 'MarkerSize', 18, 'Color', 'y');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'1')
elseif (s(y)==5)
plot(Coord(y,1),Coord(y,2),'o', 'LineWidth', 2, 'MarkerSize', 18, 'Color', 'm');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'14')
elseif (s(y)==6)
plot(Coord(y,1),Coord(y,2),'o', 'LineWidth', 2, 'MarkerSize', 18, 'Color', 'c');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'10')
elseif (s(y)==7)
plot(Coord(y,1),Coord(y,2),'o', 'LineWidth', 2, 'MarkerSize', 18, 'Color', 'k');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'6')
elseif (s(y)==8)
plot(Coord(y,1),Coord(y,2),'o', 'LineWidth', 2, 'MarkerSize', 18, 'Color', [1 0.4 0.6]);
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'2')
elseif (s(y)==9)
plot(Coord(y,1),Coord(y,2),'o', 'LineWidth', 2, 'MarkerSize', 18, 'Color', 'r');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'15')
elseif (s(y)==10)
plot(Coord(y,1),Coord(y,2),'o', 'LineWidth', 2, 'MarkerSize', 18, 'Color', 'b');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'11')
elseif (s(y)==11)
plot(Coord(y,1),Coord(y,2),'o', 'LineWidth', 2, 'MarkerSize', 18, 'Color', 'g');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'7')
elseif (s(y)==12)
plot(Coord(y,1),Coord(y,2),'o', 'LineWidth', 2, 'MarkerSize', 18, 'Color', 'y');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'3')
elseif (s(y)==13)
plot(Coord(y,1),Coord(y,2),'o', 'LineWidth', 2, 'MarkerSize', 18, 'Color', 'm');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'16')
elseif (s(y)==14)
plot(Coord(y,1),Coord(y,2),'o', 'LineWidth', 2, 'MarkerSize', 18, 'Color', 'c');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'12')
elseif (s(y)==15)
plot(Coord(y,1),Coord(y,2),'o', 'LineWidth', 2, 'MarkerSize', 18, 'Color', 'k');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'8')
elseif (s(y)==16)
plot(Coord(y,1),Coord(y,2),'o', 'LineWidth', 2, 'MarkerSize', 18, 'Color', [1 0.4 0.6]);
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'4')
end
% if training data
% plot color and number at position of dataset depending on label
elseif (label(y) == 1)
plot(Coord(y,1),Coord(y,2),'s', 'LineWidth', 2, 'MarkerSize', 20,'Color','r');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'13')
elseif(label(y) == 2)
plot(Coord(y,1),Coord(y,2),'s', 'LineWidth', 2, 'MarkerSize', 20,'Color','b');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'9')
elseif(label(y) == 3)
plot(Coord(y,1),Coord(y,2),'s', 'LineWidth', 2, 'MarkerSize', 20,'Color','g');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'5')
elseif(label(y) == 4)
plot(Coord(y,1),Coord(y,2),'s', 'LineWidth', 2, 'MarkerSize', 20,'Color','y');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'1')
elseif(label(y) == 5)
plot(Coord(y,1),Coord(y,2),'s', 'LineWidth', 2, 'MarkerSize', 20,'Color','m');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'14')
elseif(label(y) == 6)
plot(Coord(y,1),Coord(y,2),'s', 'LineWidth', 2, 'MarkerSize', 20,'Color','c');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'10')
elseif(label(y) == 7)
plot(Coord(y,1),Coord(y,2),'s', 'LineWidth', 2, 'MarkerSize', 20,'Color','k');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'6')
elseif(label(y) == 8)
plot(Coord(y,1),Coord(y,2),'s', 'LineWidth', 2, 'MarkerSize', 20,'Color',[1 0.4 0.6]);
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'2')
elseif(label(y) == 9)
plot(Coord(y,1),Coord(y,2),'s', 'LineWidth', 2, 'MarkerSize', 20,'Color','r');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'15')
elseif(label(y) == 10)
plot(Coord(y,1),Coord(y,2),'s', 'LineWidth', 2, 'MarkerSize', 20,'Color','b');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'11')
elseif(label(y) == 11)
plot(Coord(y,1),Coord(y,2),'s', 'LineWidth', 2, 'MarkerSize', 20,'Color','g');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'7')
elseif(label(y) == 12)
plot(Coord(y,1),Coord(y,2),'s', 'LineWidth', 2, 'MarkerSize', 20,'Color','y');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'3')
elseif(label(y) == 13)
plot(Coord(y,1),Coord(y,2),'s', 'LineWidth', 2, 'MarkerSize', 20,'Color','m');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'16')
elseif(label(y) == 14)
plot(Coord(y,1),Coord(y,2),'s', 'LineWidth', 2, 'MarkerSize', 20,'Color','c');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'12')
elseif(label(y) == 15)
plot(Coord(y,1),Coord(y,2),'s', 'LineWidth', 2, 'MarkerSize', 20,'Color','k');
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'8')
elseif(label(y) == 16)
plot(Coord(y,1),Coord(y,2),'s', 'LineWidth', 2, 'MarkerSize', 20,'Color',[1 0.4 0.6]);
text((Coord(y,1)-xoffset),(Coord(y,2)+yoffset),'4')
end
end
% END PLOT RESULTS
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%