LDA Mallet implementation on Design Discussions on StackOverflow [article]

Rohith Pudari, Roshan Lasrado, Dave Cheng
2020 Zenodo  
Abstract: Identifying various software design related topics that are frequently questioned by developers can help in understanding the areas that the developers often find challenging and those that require more research/educational efforts. In this paper, we conduct a study of the StackOverflow (SO) design-related dataset to identify the various categories of design-related questions asked and the challenges faced by the developers. We replicate the methods used by a previous study (Bangash
more » ... al., 2019) for the domain of Software Design. First, we identify the list of design-related topics using the Latent Dirichlet Allocation (LDA) algorithm for topic modelling and utilize inductive coding to label the topics. We then perform a qualitative analysis of the various identified design-related topics to glean the kind of challenges faced by the developers. We also study whether the topics identified by the LDA algorithm could be used for tagging design-related StackOverflow posts.
doi:10.5281/zenodo.4314692 fatcat:pnr2rfescbe5fhlj76jzchmq5u