Extracted Summary Based Recommendation System for Indian Legal Documents

摘要

Law practitioners of a Common Law legal system are often tasked with finding legal documents similar to the case at hand. However, Indian legal case documents are verbose and unstructured. Thus identifying a concise, summarised version of the case is a non-trivial task. This paper proposes a novel framework in order identify those paragraphs of the case document that contribute to its summary, and use them to retrieve other similar documents. A dataset containing explicitly labelled summary paragraphs of Indian Supreme Court documents dated prior to the 1970s has been exploited to train a Support Vector Classifier in order to achieve this goal. The model discussed in this paper has shown promising results, with a high accuracy. Using only the extracted summary to retrieve similar documents shows better performance in terms of time and space complexity as compared to considering the document as a whole.