static double[] similarity(LatentDirichletAllocation lda0, LatentDirichletAllocation lda1) {
int numTopics = lda0.numTopics();
int numPairs = numTopics * (numTopics - 1);
@SuppressWarnings({"unchecked", "rawtypes"}) // ok given use w. erasure
ScoredObject<int[]>[] pairs = (ScoredObject<int[]>[]) new ScoredObject[numPairs];
int pos = 0;
for (int i = 0; i < numTopics; ++i) {
for (int j = 0; j < numTopics; ++j) {
if (i == j) continue;
double divergence =
Statistics.symmetrizedKlDivergence(
lda0.wordProbabilities(i), lda1.wordProbabilities(j));
pairs[pos++] = new ScoredObject<int[]>(new int[] {i, j}, divergence);
}
}
Arrays.sort(pairs, ScoredObject.comparator());
boolean[] taken0 = new boolean[numTopics];
Arrays.fill(taken0, false);
boolean[] taken1 = new boolean[numTopics];
Arrays.fill(taken1, false);
double[] scores = new double[numTopics];
int scorePos = 0;
for (pos = 0; pos < numPairs && scorePos < numTopics; ++pos) {
int[] pair = pairs[pos].getObject();
if (!taken0[pair[0]] && !taken1[pair[1]]) {
taken0[pair[0]] = true;
taken1[pair[1]] = true;
scores[scorePos++] = pairs[pos].score();
}
}
return scores;
}
public static void main(String[] args) throws Exception {
int minTokenCount = 5;
File corpusFile = new File(args[0]);
CharSequence[] articleTexts = LdaWormbase.readCorpus(corpusFile);
SymbolTable symbolTable = new MapSymbolTable();
TokenizerFactory tokenizerFactory = LdaWormbase.WORMBASE_TOKENIZER_FACTORY;
int[][] docTokens =
LatentDirichletAllocation.tokenizeDocuments(
articleTexts, tokenizerFactory, symbolTable, minTokenCount);
LdaRunnable runnable1 =
new LdaRunnable(docTokens, new LdaReportingHandler(symbolTable), new Random());
LdaRunnable runnable2 =
new LdaRunnable(docTokens, new LdaReportingHandler(symbolTable), new Random());
Thread thread1 = new Thread(runnable1);
Thread thread2 = new Thread(runnable2);
thread1.start();
thread2.start();
thread1.join();
thread2.join();
LatentDirichletAllocation lda0 = runnable1.mLda;
LatentDirichletAllocation lda1 = runnable2.mLda;
System.out.println("\nComputing Greedy Aligned Symmetrized KL Divergences");
double[] scores = similarity(lda0, lda1);
for (int i = 0; i < scores.length; ++i) System.out.printf("%4d %15.3f\n", i, scores[i]);
}
static LatentDirichletAllocation sample(
int[][] docTokens,
ObjectHandler<LatentDirichletAllocation.GibbsSample> handler,
Random random) {
short numTopics = 50;
double topicPrior = 0.1;
double wordPrior = 0.01;
int burninEpochs = 0;
int sampleLag = 1;
int numSamples = 2000;
LatentDirichletAllocation.GibbsSample sample =
LatentDirichletAllocation.gibbsSampler(
docTokens,
numTopics,
topicPrior,
wordPrior,
burninEpochs,
sampleLag,
numSamples,
random,
handler);
return sample.lda();
}
