maziyarpanahi

Phrases were extracted by Stanford CoreNLP/Spark 2.2 (6minutes) from English Wikipeida (+5 million pages)

+---------------------------+-----+                                                                                                                                                                                                                                   [441/9895]
|value                      |count|
+---------------------------+-----+
|square miles               |59821|
|unique feature             |46463|
|id form                    |46101|
|administrative district    |45963|
|first time                 |41423|

maziyarpanahi

Stanford CoreNLP (Sentence splitter and POS Tagging - extract noun phrases), StopWordsRemover, TF-IDF, word2vec and OnlineLDAOptimizer

==========
Query: "cancer"
Sample: 500K abstracts
Dataset: PubMed
==========
val numTopics: Int = 50
val maxIterations: Int = 100
val vocabSize: Int = 10000

maziyarpanahi

Stanford CoreNLP (Sentence splitter and POS Tagging - extract noun phrases), StopWordsRemover, TF-IDF, word2vec and OnlineLDAOptimizer
Query: Global Warming (5000 pages)

==========Parameteres==========
val numTopics: Int = 50
val maxIterations: Int = 100
val vocabSize: Int = 10000
val minDF: Int = 1
val minTF: Int = 1
val maxItems: Int = 15

maziyarpanahi

====================
Stanford CoreNLP (Sentence splitter and POS Tagging - NN and NNS), StopWordsRemover, TF-IDF, word2vec and OnlineLDAOptimizer
Query: Global Warming (5000 pages)

==========Parameteres==========
val numTopics: Int = 50
val maxIterations: Int = 100
val vocabSize: Int = 10000
val minDF: Int = 10
val minTF: Int = 1