Using machine learning algorithms to identify named entities in legal documents: a preliminary approach

Abstract

This paper deals with accuracy and performance of var- ious machine learning algorithms in the recognition and extraction of different types of named entities such as date, organization, reg- ulation laws and person. The experiment is based on 20 judicial decision documents from European Lex site. The obtained results were proposed for the selection of the best algorithm that selects appropriate maximum entities from the legal documents. To ver- ify the performance of algorithm, obtained data from the tagging entities were compared with manual work as reference.