The paper presents an empirical study of integrating ngrams and multi-word terms into topic models, while maintaining similarities between them and words based on their component structure. First, we adapt the PLSA-SIM algorithm to the more widespread LDA model and ngrams. Then we propose a novel algorithm LDA-ITER that allows the incorporation of the most suitable ngrams into topic models. The experiments of integrating ngrams and multiword terms conducted on five text collections in different languages and domains demonstrate a significant improvement in all the metrics under consideration.
The paper describes the results of an empirical study of integrating bigram collocations and similarities between them and unigrams into topic models. First of all, we propose a novel algorithm PLSA-SIM that is a modification of the original algorithm PLSA. It incorporates bigrams and maintains relationships between unigrams and bigrams based on their component structure. Then we analyze a variety of word association measures in order to integrate top-ranked bigrams into topic models. All experiments were conducted on four text collections of different domains and languages. The experiments distinguish a subgroup of tested measures that produce topranked bigrams, which demonstrate significant improvement of topic models quality for all collections, when integrated into PLSA-SIM algorithm.
Abstract. In this paper we present the approach of introducing thesaurus knowledge into probabilistic topic models. The main idea of the approach is based on the assumption that the frequencies of semantically related words and phrases, which are met in the same texts, should be enhanced: this action leads to their larger contribution into topics found in these texts. We have conducted experiments with several thesauri and found that for improving topic models, it is useful to utilize domain-specific knowledge. If a general thesaurus, such as WordNet, is used, the thesaurus-based improvement of topic models can be achieved with excluding hyponymy relations in combined topic models.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.