Complexity Analysis of Legal Documents

Legal documents are considered one of the most difficult types of documents to write and interpret due to the fact that they use legal jargon and follow a specific legal convention. For a legal practitioner, they require documents, mainly judgements, for the proceedings in their client's case. We evaluated the approach of Named Entity Recognition on a large corpus of legal judgements and found 14 entities, such as case number, name of judge, name of court, statute, provision, etc. Through this, we found that one cannot comprehend the document completely as it is written in a complex manner. We also conduct a complexity analysis of the legal documents using a variety of linguistic features, including sentence length, word frequency, and syntactic complexity. Considering the presence of legal jargon, we defined a method to analyse the complexity on these legal documents and a suitable score to understand them. Given that there are no specific ways for defining the complexity of legal documents, we performed human evaluation and achieved the best value for the recommended approaches with the assistance of legal experts.