-
Notifications
You must be signed in to change notification settings - Fork 1
/
KmeanClustering.java
462 lines (411 loc) · 14.3 KB
/
KmeanClustering.java
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
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
//This is the main algorithm that uses prominent attributes and/or all attributes if needed
// merge them
//Author: Shehroz S. Khan
//Affiliation: University of Waterloo, Canada
//Date: May'2012
//LICENCE: Read Separate File
//About: This is the main class to generate initial modes and perform K-modes clustering
// Details of the algorithm are in the paper
package initKmean;
import java.io.BufferedWriter;
import java.io.FileWriter;
import java.io.IOException;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collections;
import java.util.Comparator;
import java.util.HashMap;
import java.util.Iterator;
import java.util.LinkedHashMap;
import java.util.LinkedList;
import java.util.List;
import java.util.Map;
import java.util.Map.Entry;
import org.apache.commons.lang3.ArrayUtils;
import org.apache.commons.math3.ml.distance.EuclideanDistance;
import org.apache.commons.math3.special.Erf;
import org.apache.commons.math3.stat.descriptive.DescriptiveStatistics;
import weka.core.DenseInstance;
import weka.core.Instance;
import weka.core.Instances;
import weka.core.converters.ConverterUtils.DataSource;
import weka.filters.Filter;
import weka.filters.unsupervised.attribute.Normalize;
import weka.filters.unsupervised.attribute.Remove;
class KmeanClustering {
private int K;
private int NN;
private int ITR_MAX;
private int [] count;
private double [][] newMeans;
private int [][] mCluster ;
//Constructor
public KmeanClustering () {
}
//Setters
public void setK(int k) {
K = k;
}
public void setNN(int nN) {
NN = nN;
}
public void setITR_MAX (int itr){
ITR_MAX=itr;
}
//Getters
public int getK() {
return K;
}
public int getNN() {
return NN;
}
public double [][] getMeans(){
return newMeans;
}
public int [][] getmCluster() {
return mCluster;
}
public int [] getObjectCountInClusters(){
return count;
}
public int getITR_MAX (){
return ITR_MAX;
}
// CCIA - Cluster Center Initialization Algorithm
public double [][] CCIA (Instances data, String outputFile) throws Exception {
System.out.println("Total Instances="+data.numInstances());
int [][] clusterString = new int [data.numInstances()][data.numAttributes()];
//Find centers corresponding to each attribute
for(int i=0;i<data.numAttributes();i++) {
System.out.println("Atr="+i);
double [] val = new double[data.numInstances()];
for (int j=0;j<data.numInstances();j++) {
val[j]=data.instance(j).value(i);
//System.out.println("j="+j+", val="+val[j]);
}
int [] str=clusterNumericAttribute(val,data);
//for (int j=0;j<str[i].length;j++)
//System.out.print(str[i][j]+" ");
//System.out.println();
int [] membership=generateClusterString(str,data);
for(int l=0;l<data.numInstances();l++) {
clusterString[l][i]=membership[l];
}
} //end for each attributes
String [] cstr = extractClusterStrings(clusterString,data);
Map<String, Integer> distinctClassStr = findUniqueClusterStrings(cstr);
double [][] initCenters = findInitialCenters(cstr,distinctClassStr, data);
return initCenters;
}
//Find Initial Centers for Kmeans clustering
public double [][] findInitialCenters(String [] cstr, Map<String, Integer> distinctClassStr, Instances data) throws Exception {
double [][] initCenters = new double [distinctClassStr.size()][data.numAttributes()];
int [] count = new int [distinctClassStr.size()];
for(int i=0;i<cstr.length;i++) {
int j=0;
Iterator it = distinctClassStr.entrySet().iterator();
while (it.hasNext()) {
Map.Entry pairs = (Map.Entry)it.next();
//System.out.println(pairs.getKey() + " = " + pairs.getValue()+" -->"+pairs.getKey().toString().equals(cstr[i]));
//Store all strings
//topclusterString[i]=pairs.getKey().toString();
if(pairs.getKey().toString().equals(cstr[i])) {
for(int k=0;k<data.numAttributes();k++)
initCenters[j][k]+=data.instance(i).value(k);
count[j]++;
break;
}
j++;
}
}
for (int i=0;i<distinctClassStr.size();i++) {
for (int j=0;j<data.numAttributes();j++) {
initCenters[i][j]=initCenters[i][j]/count[i];
//System.out.print(initCenters[i][j]+" ");
}
//System.out.println();
}
if (distinctClassStr.size()==getK())
return initCenters;
else
return initCenters = MergeDBMSDC(initCenters,distinctClassStr,data);
}
//Merge DBMSDC algorithm to merge similar centers
public double [][] MergeDBMSDC(double [][] initCenters, Map<String, Integer> distinctClassStr , Instances data) throws Exception {
double [][] centers = new double [getK()][data.numAttributes()];
int [] B = new int [distinctClassStr.size()];
for (int i=0;i<distinctClassStr.size();i++)
B[i]=i;
int L;
for(L=0;L<getK()-1;L++) {
if(B.length <= getNN())
throw new Exception ("\n***ATTENTION*** The number of nearest neighbours are more than the centers. "
+ "Consider reducing the number of nearest numbers using the function setNN() in testKmeans.java");
double [] R = new double [B.length];
for (int i=0;i<B.length;i++) {
double [] distance = new double [B.length];
for (int j=0;j<B.length;j++) {
EuclideanDistance ed = new EuclideanDistance();
distance[j]=ed.compute(initCenters[i], initCenters[j]);
}
double [] sort= Arrays.copyOf(distance, distance.length);
Arrays.sort(sort);
R[i]=sort[getNN()];
}
DescriptiveStatistics stat = new DescriptiveStatistics(R);
double minR = stat.getMin();
int index=0;
for (int i=0;i<R.length;i++) {
if(R[i]==minR) {
index=i;
break;
}
}
ArrayList<Integer> S = new ArrayList<Integer>();
for (int i=0;i<B.length;i++) {
EuclideanDistance ed = new EuclideanDistance();
double dist = ed.compute(initCenters[index], initCenters[i]);
if (dist < 1.5*minR) {
S.add(B[i]);
B=ArrayUtils.removeAllOccurences(B, B[i]);
}
}
double [] temp = new double [data.numAttributes()];
for (int i=0;i<S.size();i++) {
for(int j=0;j<data.numAttributes();j++){
temp[j]+=initCenters[S.get(i)][j]/S.size();
}
}
centers[L]=temp;
}
//Merge the remaining centers as the final center
double [] temp = new double [data.numAttributes()];
for (int i=0;i<B.length;i++) {
for(int j=0;j<data.numAttributes();j++) {
temp[j]=initCenters[B[i]][j]/B.length;
}
}
centers[L]=temp;
// for(int j=0;j<data.numAttributes();j++)
// centers[L][j]=initCenters[B[0]][j];
// System.out.println("Desired number of "+getK()+" clusters are generated");
return centers;
}
//Extract clustering strings for the whole data
public String [] extractClusterStrings(int [][] clusterString, Instances data) {
//Convert numeric class string to character strings
String [] cstr = new String [data.numInstances()];
for(int i=0;i<data.numInstances();i++) {
cstr[i]="";
for(int j=0; j<data.numAttributes()-1;j++){
cstr[i]+=clusterString[i][j]+",";
//System.out.print(clusterString[i][j]+", ");
}
cstr[i]+=clusterString[i][data.numAttributes()-1];
//System.out.println(cstr[i]);
//System.out.println();
}
return cstr;
}
//Find unique cluster strings
public Map<String, Integer> findUniqueClusterStrings(String [] cstr) {
//Find distinct class strings
Map<String, Integer> distinctClassStr = distinctAttributes(cstr);
//System.out.println(distinctClassStr);
System.out.println("\nDistinct Cluster Strings="+distinctClassStr.size()+"\n");
return distinctClassStr;
}
//Sort a map by value
public LinkedHashMap sortByValue(Map<String, Integer> map) {
List list = new LinkedList(map.entrySet());
Collections.sort(list, new Comparator() {
public int compare(Object o1, Object o2) {
return ((Comparable) ((Map.Entry) (o1)).getValue())
.compareTo(((Map.Entry) (o2)).getValue());
}
});
Map result = new LinkedHashMap();
for (Iterator it = list.iterator(); it.hasNext();) {
Map.Entry entry = (Map.Entry)it.next();
result.put(entry.getKey(), entry.getValue());
}
return (LinkedHashMap) result;
} //end sortByValue
//Generate cluster strings for each attribute
public int [] generateClusterString(int [] str, Instances data) {
//Find new centers corresponding to this attributes cluster allotments
//Allot data objects based on cluster allotments
double [][] clust = new double [getK()][data.numAttributes()];
int [] count = new int [getK()];
for (int i=0;i<str.length;i++) {
for (int j=0;j<data.numAttributes();j++) {
clust[str[i]][j]+=data.instance(i).value(j);
}
count[str[i]]++;
}
for (int i=0;i<getK();i++) {
for (int j=0;j<data.numAttributes();j++) {
clust[i][j]=clust[i][j]/count[i];
}
}
System.out.println("\nUsing centers derived from this attribute");
//Perform Kmeans with these initial centers
int [] membership = KMeansClustering(data,clust,getK(),data.numAttributes());
return membership;
}
public void writeOutput (double [][] initCenters, String outputFile) throws IOException {
BufferedWriter output = new BufferedWriter(new FileWriter(outputFile));
for(int i=0;i<initCenters.length;i++) {
for (int j=0;j<initCenters[i].length;j++) {
output.write(initCenters[i][j]+",");
}
output.newLine();
}
output.close();
}
//Cluster numeric attribute
public int [] clusterNumericAttribute(double [] attrib,Instances data) {
double [][] xs = new double [getK()][1];
// Normalize attribute values
DescriptiveStatistics stats = new DescriptiveStatistics(attrib);
double mean = stats.getMean();
double sd = stats.getStandardDeviation();
//System.out.println("m="+mean+" sd="+sd);
for (int i=0; i<getK();i++) {
double percentile=(double)(2*(i+1)-1)/(2*getK());
double z = Math.sqrt(2) * Erf.erfcInv(2*percentile); //https://stats.stackexchange.com/questions/71788/percentile-to-z-score-in-php-or-java
xs[i][0]=z*sd+mean;
//System.out.println("p="+percentile+ " z="+z+" xs["+i+"]="+xs[i][0]);
}
//Convert 'this' attribute to Weka Instances
Instances ad = new Instances(data,data.numInstances());
for (int i=0;i<attrib.length;i++) {
Instance ins = new DenseInstance(1);
ins.setValue(0, attrib[i]);
ad.add(i, ins);
}
//System.out.println(ad);;
//Perform Kmeans on 'this' attribute using xs as initial centers
int [] membership = KMeansClustering(ad,xs,getK(),1);
return membership;
}
// K-means Clustering
public int[] KMeansClustering (Instances data, double [][] means, int K,int numAttr){
int membership [] = new int [data.numInstances()];
for(int itr=0;itr<ITR_MAX;itr++) {
count = new int [K];
mCluster = new int [K][data.numInstances()];
System.out.println("------------------ITR="+(itr+1)+"-----------------------");
//Partition data based on initial means
for(int i=0;i<data.numInstances();i++){
double distance [] = new double [K];
for(int j=0;j<K;j++){
double [] attr = new double [numAttr];
double [] mattr = new double [numAttr];
for(int k=0;k<numAttr;k++) {
attr[k] = data.instance(i).value(k);
mattr[k]=means[j][k];
}
//distance[j] = computeHammingDistance(str,modes[j], data.numAttributes());
EuclideanDistance ed = new EuclideanDistance();
distance[j]=ed.compute(attr, mattr);
} //end for j
//Find membership of instances to clusters w.r.t. hamming distance
//for(int j=0;j<K;j++)
//System.out.print(distance[j]+" ");
membership[i] = findClusterMembership(distance);
//System.out.println("m="+membership[i]+" c="+count[membership[i]]);
mCluster[membership[i]][count[membership[i]]]=i;
count[membership[i]]++;
} //end for i
//for(int i=0;i<K;i++)
// System.out.print("cluster["+i+"]="+count[i]+" ");
//System.out.println();
//Allocate instances to K clusters
newMeans = new double [K][numAttr];
for(int i=0;i<K;i++){
for(int k=0;k<numAttr;k++) {
double [] val = new double [count[i]];
for(int j=0;j<count[i];j++){
//System.out.print(data.instance(mCluster[i][j]).stringValue(k)+" ");
val[j] = data.instance(mCluster[i][j]).value(k);
} //end for j
DescriptiveStatistics stats = new DescriptiveStatistics(val);
newMeans[i][k]=stats.getMean();
//System.out.print(newMeans[i][k]+" ");
} //end for k
//System.out.println();
} //end for i
//System.out.println();
//Check termination condition
//System.out.println("D="+data.numAttributes());
int flag=1;
for(int i=0;i<K;i++){
for(int j=0;j<numAttr;j++){
if(means[i][j]==newMeans[i][j]) flag*=1;
else flag=0;
}
}
if(flag==0) means=newMeans;
else if (flag==1) break;
//Count empty clusters
int emptyCluster=0;
for (int i=0;i<K;i++) {
if (count[i]==0) {
emptyCluster++;
}
}
//Reduce the number of clusters K
K=K-emptyCluster;
} //end for itr
return membership;
}
//Find cluster membership of a data object
public int findClusterMembership(double[] distance) {
double [] temp = new double [distance.length];
for(int i=0;i<temp.length;i++) temp[i]=distance[i];
Arrays.sort(temp);
int i;
for(i=0;i<distance.length;i++){
if(temp[0]==distance[i])
break;
}
return i;
}
// Compute Distinct Attribute Values
public Map<String, Integer> distinctAttributes (String [] args){
Map<String, Integer> m = new HashMap<String, Integer>();
for (String a : args) {
Integer freq = m.get(a);
m.put(a, (freq == null) ? 1 : freq + 1);
}
//System.out.println(m.size() + " distinct words:" + m);
return m;
}
//Read input file
public Instances readInputFile(String inputCSVfile) throws Exception {
DataSource source = new DataSource(inputCSVfile);
Instances data = source.getDataSet();
// setting class attribute if the data format does not provide this information
if (data.classIndex() == -1)
data.setClassIndex(data.numAttributes() - 1);
//System.out.println(data);
return data;
} //end readInputFile
//Remove class attribute
public Instances removeClassAttribute(Instances data) throws Exception {
// generate data for clusterer (w/o class)
Remove filter = new Remove();
filter.setAttributeIndices("" + (data.classIndex() + 1));
filter.setInputFormat(data);
Instances dataClusterer = Filter.useFilter(data, filter);
return dataClusterer;
} //end for removeClassAttribute()
//Normalize data
public Instances normalizeData(Instances data) throws Exception {
//normalize
Normalize normalizeFilter = new Normalize();
normalizeFilter.setInputFormat(data);
return data = Filter.useFilter(data, normalizeFilter);
}
}