Incorporating structural topic modeling into short text analysis

Abstract The past few decades have seen the rapid development of topic modeling. So far, research has been more concerned with determining ideal number topics or meaningful clustering words than applying modeling techniques to evaluate linguistic theories. This study proposes Structural Topic Model (STM)-led framework facilitate interpretation results and standardize text analysis. STM encompasses various model training mechanisms, thereby requiring systematic designs properly combine language studies. “Structural” in refers inclusion metadata structure. Unlike corpus-based keyness approach, can capture contextual cues meta-information for topical results. Besides, make cross-corpora comparisons via contrast, a challenging task corpus-driven related models such as Biterm (BTM). Stylistic variations song lyrics are taken an illustration show how use suggested delve into theory proposed by Pennebaker (2013) . iterable paradigm clarify pronouns affect style distinction. We believe STM-led shed light on analysis conducting reproducible comparison short texts.