Review of Trends in Topic Modeling Techniques, Tools, Inference Algorithms and Applications
Christine K. Mulunda, Peter Waiganjo Wagacha, Lawrence Muchemi · 2018
Topic modeling algorithms discover latent abstract topics in large volumes of data. The choice of inference algorithm determines the kind of solution the topic models will provide. Inference algorithms aim at fast convergence of topics and stability of solutions with respect to initial parameters. The problem area identified for a topic model forms the basis of selecting the inference algorithm. This study reviewed 85 articles from various journals, classified them as reviews, techniques, tools, inference algorithms for topic models and application areas, among others. Scalability, computation speed and dynamism of topic models are necessary for robust solutions in analysis of datasets that are voluminous, varied and that scales up very fast with time.