Topic Modelling Supreme Court Case Decisions using Latent Dirichlet Allocation

Abstract

Extracting dominant topics from a bulk amount of Philippine Supreme Court case decisions can be time consuming and intensive. Furthermore, being able to find relevant topics automatically without close reading can aid legal researchers, law practitioners, and historians in studying and analyzing judicial behaviors of the high court. The six topic models generated in this study shows to have an average score of 86.67% in determining the dominant topic of case decisions via modified topic intrusion experiment.