Indexing and searching of judicial precedents using automatic summarization

Abstract

With the aim of democratizing access to justice, the Colombian legal system has recognized the importance of judicial precedent. Judicial precedent allows citizens to request judicial decisions based on previous sentences (court rulings). Some tools are provided to search these sentences. These tools are based on keywords and therefore the search may obtain many non-relevant results. The low efficacy of these tools may make it difficult to find the right sentence. Recently, an approach for sentences searching in the Colombian context was presented. This approach explores the full content of the sentences, which leads to high processing times. One solution to this problem may be to generate automatic summaries before performing the search. This paper presents a comparative analysis of algorithms for automatic summaries. The experimental evaluation shows promising results.