Evaluating Document Analysis with kNN Based Approaches in Judicial Offices of Bangladesh

Abstract

In this contemporary era of artificial intelligence, machine learning (ML) algorithms are getting significant attention for the analysis of textual analysis. In recent years, operational improvement in different corporate sectors of Bangladesh are achieved by implementing digitization of the process flow instead of using manual paper trails in offices. Nowadays, judicial sectors are included into sate wide digitalization process by archiving the judiciary records. Despite such improvement, autonomic categorizing of documents using textual analysis is not seen in labeling the correct class of a judicial document. In fact, officers spend lots of time in manual labeling of court related document. In our present investigation, we approached a textual analysis tool that can initiate towards the major solution for solving the manual categorization problem within the judicial sector of Bangladesh. Our objective is to label a normalized text document by implementing ML algorithm into suitable class in terms of the case type. In addition, grammatical analysis of English documents is integrated by the natural language processing (NLP) techniques as well as the filtering of feature sets by TF-IDF based term weighting scheme. The outcomes show the important impacts of NLP techniques for generating useful training data in KNN classification algorithm for the categorization of English documents in Bangladeshi judiciary sector.