Topic modelling is a state-of-the-art technique to understand, categorize and summarise the text and is beneficial to discover the hidden themes in text collections. The existing topic modelling approaches pay less or no attention to capturing the semantics of words. Hence, meaningless topics are generated. This research addresses the main problem of existing topic modelling approaches by introducing two semantic-based topic generation approaches. The thesis has made main contributions to topic modelling and text mining domains by introducing semantic-based topic representation, semantic topic model and an ambiguity handling approach. The research outcomes are beneficial for many text mining applications.
原文链接
分享
引用
