A Comparative Study of Summarization Algorithms Applied to Legal Case Judgments

• Published in: Advances in Information Retrieval pp. 413 - 428
• Main Authors: Bhattacharya, Paheli, Hiware, Kaustubh, Rajgaria, Subham, Pochhi, Nilay, Ghosh, Kripabandhu, Ghosh, Saptarshi
• Format: Book Chapter
• Language: English
• Published: Cham Springer International Publishing
• Series: Lecture Notes in Computer Science
• Subjects: Legal case judgment; Summarization; Supervised; Unsupervised
• ISBN: 3030157113; 9783030157111
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-030-15712-8_27

Summarization of legal case judgments is an important problem because the huge length and complexity of such documents make them difficult to read as a whole. Many summarization algorithms have been proposed till date, both for general text documents and a few specifically targeted to summarizing legal documents of various countries. However, to our knowledge, there has not been any systematic comparison of the performances of different algorithms in summarizing legal case documents. In this paper, we perform the first such systematic comparison of summarization algorithms applied to legal judgments. We experiment on a large set of Indian Supreme Court judgments, and a large variety of summarization algorithms including both unsupervised and supervised ones. We assess how well domain-independent summarization approaches perform on legal case judgments, and how approaches specifically designed for legal case documents of other countries (e.g., Canada, Australia) generalize to Indian Supreme Court documents. Apart from quantitatively evaluating summaries by comparing with gold standard summaries, we also give important qualitative insights on the performance of different algorithms from the perspective of a law expert.