Abstract | ||
---|---|---|
This paper introduces the novel task of topic coherence evaluation, whereby a set of words, as generated by a topic model, is rated for coherence or interpretability. We apply a range of topic scoring models to the evaluation task, drawing on WordNet, Wikipedia and the Google search engine, and existing research on lexical similarity/relatedness. In comparison with human scores for a set of learned topics over two distinct datasets, we show a simple co-occurrence measure based on pointwise mutual information over Wikipedia data is able to achieve results for the task at or nearing the level of inter-annotator correlation, and that other Wikipedia-based lexical relatedness methods also achieve strong results. Google produces strong, if less consistent, results, while our results over WordNet are patchy at best. |
Year | Venue | Keywords |
---|---|---|
2010 | north american chapter of the association for computational linguistics | strong result,topic coherence evaluation,lexical similarity,google search engine,automatic evaluation,topic model,evaluation task,wikipedia-based lexical relatedness method,wikipedia data,distinct datasets,novel task |
Field | DocType | ISBN |
Normalized Google distance,Lexical similarity,Computer science,Natural language processing,Artificial intelligence,WordNet,Interpretability,Search engine,Information retrieval,Coherence (physics),Topic model,Pointwise mutual information,Machine learning | Conference | 1-932432-65-5 |
Citations | PageRank | References |
203 | 7.42 | 27 |
Authors | ||
4 |
Name | Order | Citations | PageRank |
---|---|---|---|
David Newman | 1 | 1319 | 73.72 |
Jey Han Lau | 2 | 660 | 36.88 |
Karl Grieser | 3 | 295 | 11.68 |
Timothy Baldwin | 4 | 1767 | 116.85 |