1
我有一個LDA模型，運行在語料庫大小爲12,054個文檔，語義大小爲9,681個字和60個集羣。我試圖通過調用.topicDistributions（）或.javaTopicDistributions（）來獲取文檔的主題分佈。這兩種方法都會在文檔上返回一個主題分佈的rdd。根據我的理解，行數應該是文檔數量，列數應該是主題數量。但是，在調用topicDistributions（）之後，當我計算rdd時，我得到的計數爲11,665（比傳遞給模型的文檔數少）？每個文檔都有正確數量的主題（60）。爲什麼是這樣？Spark 1.4 Mllib LDA topicDistributions（）返回錯誤的文檔數
這裏的演示： http://spark.apache.org/docs/latest/mllib-clustering.html
下面的代碼：
enter code here
//parse tf vectors from corpus
JavaRDD<Vector> parsedData = data.map(
new Function<String, Vector>() {
public Vector call(String s) {
s = s.substring(1, s.length()-1);
String[] sarray = s.trim().split(",");
double[] values = new double[sarray.length];
for (int i = 0; i < sarray.length; i++)
{
values[i] = Double.parseDouble(sarray[i]);
}
return Vectors.dense(values);
}
);
System.out.println(parsedData.count()) //prints 12,054
// Index documents with unique IDs
JavaPairRDD<Long, Vector> corpus = JavaPairRDD.fromJavaRDD(parsedData.zipWithIndex().map(
new Function<Tuple2<Vector, Long>, Tuple2<Long, Vector>>() {
public Tuple2<Long, Vector> call(Tuple2<Vector, Long> doc_id) {
return doc_id.swap();
}
}
));
System.out.println(corpus.count()) //prints 12,054
LDA lda = new LDA()
LDAModel ldaModel = lda.setK(k.intValue()).run(corpus);
RDD<scala.Tuple2<Object,Vector>> topic_dist_over_docs = ((DistributedLDAModel) ldaModel).topicDistributions();
System.out.println(topic_dist_over_docs.count()) //prints 11,655 ???
JavaPairRDD<Long,Vector> topic_dist_over_docs2 = ((DistributedLDAModel) ldaModel).javaTopicDistributions();
System.out.println(topic_dist_over_docs2.count()) //also prints 11,655 ???