ealdent-lda-ruby 0.1.4 → 0.1.5

data/lib/lda-inference.c

@@ -469,7 +469,21 @@ void run_quiet_em(char* start, corpus* corpus) {
 	if (strcmp(start, "seeded")==0) {
 		model = new_lda_model(corpus->num_terms, NTOPICS);
 		ss = new_lda_suffstats(model);
-		corpus_initialize_ss(ss, model, corpus);
+		if (VERBOSE) {
+		    corpus_initialize_ss(ss, model, corpus);
+	    } else {
+            quiet_corpus_initialize_ss(ss, model, corpus);
+	    }
+		if (VERBOSE) {
+		    lda_mle(model, ss, 0);
+		} else {
+		    quiet_lda_mle(model, ss, 0);
+		}
+		model->alpha = INITIAL_ALPHA;
+	} else if (strcmp(start, "fixed")==0) {
+	    model = new_lda_model(corpus->num_terms, NTOPICS);
+		ss = new_lda_suffstats(model);
+		corpus_initialize_fixed_ss(ss, model, corpus);
 		if (VERBOSE) {
 		    lda_mle(model, ss, 0);
 		} else {

data/lib/lda-model.c

@@ -19,6 +19,7 @@
 
 #include "lda-model.h"
 
+
 /*
 * compute MLE lda model from sufficient statistics
 	*
@@ -70,7 +71,7 @@ void quiet_lda_mle(lda_model* model, lda_suffstats* ss, int estimate_alpha) {
 	}
 	if (estimate_alpha == 1)
 	{
-		model->alpha = opt_alpha(ss->alpha_suffstats,
+		model->alpha = quiet_opt_alpha(ss->alpha_suffstats,
 			ss->num_docs,
 			model->num_topics);
 
@@ -168,6 +169,52 @@ void corpus_initialize_ss(lda_suffstats* ss, lda_model* model, corpus* c)
 	}
 }
 
+void quiet_corpus_initialize_ss(lda_suffstats* ss, lda_model* model, corpus* c)
+{
+	int num_topics = model->num_topics;
+	int i, k, d, n;
+	document* doc;
+
+	for (k = 0; k < num_topics; k++)
+	{
+		for (i = 0; i < NUM_INIT; i++)
+		{
+			d = floor(myrand() * c->num_docs);
+			doc = &(c->docs[d]);
+			for (n = 0; n < doc->length; n++)
+			{
+				ss->class_word[k][doc->words[n]] += doc->counts[n];
+			}
+		}
+		for (n = 0; n < model->num_terms; n++)
+		{
+			ss->class_word[k][n] += 1.0;
+			ss->class_total[k] = ss->class_total[k] + ss->class_word[k][n];
+		}
+	}
+}
+
+
+/*
+ * Use the first num_topics documents of the corpus as the seeds.  If num_topics > num_docs, results might be hairy.
+ */
+void corpus_initialize_fixed_ss(lda_suffstats* ss, lda_model* model, corpus* c) {
+    int num_topics = MIN(model->num_topics, c->num_docs);
+    int k, n;
+    document* doc;
+    
+    for (k = 0; k < num_topics; k++) {
+        doc = &(c->docs[k]);
+        for (n = 0; n < doc->length; n++) {
+            ss->class_word[k][doc->words[n]] += doc->counts[n];
+        }
+        for (n = 0; n < model->num_terms; n++) {
+            ss->class_word[k][n] += 1.0;
+            ss->class_total[k] = ss->class_total[k] + ss->class_word[k][n];
+        }
+    }
+}
+
 /*
 * allocate new lda model
 	*

data/lib/lda-model.h

@@ -10,12 +10,15 @@
 
 #define myrand() (double) (((unsigned long) randomMT()) / 4294967296.)
 #define NUM_INIT 1
+#define MIN(A,B) (int)((A > B) ? (B) : (A))
 
 void free_lda_model(lda_model*);
 void save_lda_model(lda_model*, char*);
 lda_model* new_lda_model(int, int);
 lda_suffstats* new_lda_suffstats(lda_model* model);
 void corpus_initialize_ss(lda_suffstats* ss, lda_model* model, corpus* c);
+void quiet_corpus_initialize_ss(lda_suffstats* ss, lda_model* model, corpus* c);
+void corpus_initialize_fixed_ss(lda_suffstats* ss, lda_model* model, corpus* c);
 void random_initialize_ss(lda_suffstats* ss, lda_model* model);
 void zero_initialize_ss(lda_suffstats* ss, lda_model* model);
 void lda_mle(lda_model* model, lda_suffstats* ss, int estimate_alpha);

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: ealdent-lda-ruby
 version: !ruby/object:Gem::Version 
-  version: 0.1.4
+  version: 0.1.5
 platform: ruby
 authors: 
 - Jason M. Adams
@@ -10,7 +10,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2008-11-20 00:00:00 -08:00
+date: 2008-11-21 00:00:00 -08:00
 default_executable: 
 dependencies: []