Abstractive Text Summarization of Judicial Cases in Supreme Court of Pakistan

Abstract

Due to the large quantity of legal data availability on the internet, and other resources, it is vital for the research groups to carry broad research in the field of legal text processing, which can assist us make sense out of the huge quantity of obtainable data. This data expansion has forced the necessity to build systems that can help legal professionals as well as common citizens get important legal information with very little work. Legal document sum marization is one of the most vital areas in legal domains. In this research, we apply and evaluate the performance of the hybrid text summarization and simplification (HTSS) algorithm on the dataset of court hearings of the Supreme Court of Pakistan. The results showed that the system-generated outlines are not very accurate despite having attained good scores from evaluation metrices like ROUGE, SARI and CSS.