Title: Ulysses-RFSQ: A Novel Method to Improve Legal Information Retrieval Based on Relevance Feedback
Abstract: Obtaining relevant legal documents fast, from very large datasets, is essential for the proper functioning of justice and legislative institutions. Nevertheless, legacy systems currently used by these institutions in Brazil are usually outdated, requiring a large deal of manual work. Legal Information Retrieval focuses on building new methods to deal with the large amount of legal texts, allowing the retrieval of relevant information from them. Relevance Feedback, an important aspect of information retrieval systems, uses the information given by the user to improve the document retrieval for a specific request. However, expanding its use to other queries is a difficult task. A possible approach is to use Relevance Feedback information from past, similar queries. In this paper, we propose Ulysses-RFSQ, a method based on this approach which gives a bonus for the documents marked as relevant for similar queries, and, through this bonus, updates the ranking created by a relevance score based Information Retrieval algorithm, which measures the similarity between the query text and the documents to be retrieved. Due to the lack of available datasets containing relevance information for similar queries, we used a corpus of legislative requests from the Brazilian Chamber of Deputies, which are in most cases redundant, allowing the assessment of the proposed method. According to the experimental results, adding the Relevance Feedback bonus to the documents score improved the Recall@20 of a BM25 algorithm by almost 3% in the legal dataset used.
Publication Year: 2022
Publication Date: 2022-01-01
Language: en
Type: book-chapter
Indexed In: ['crossref']
Access and Citation
Cited By Count: 2
AI Researcher Chatbot
Get quick answers to your questions about the article from our AI researcher chatbot