Pages: (32-46 )
Abstract
The automated prediction of court verdicts and sentencing has been a herculean task for individuals who have court cases and legal practitioners who handle such cases in the court. The existing models are not robust enough in predicting court cases and sentencing status if the person is convicted. There are issues of misclassification leading to wrong prediction of legal case briefs and collocation which erode the semantics of summarized case files. In this paper, an improved hybrid text mining and summarization model for predicting legal cases and its sentence status if convicted was developed. Object Oriented Analysis Methodology was used in the design and python programming language was used in building the deep learning model and in the development of the application. The results from application were able to predict legal cases files according to crime types and sentence status which shows that the system performed tremendously well and has the potential of assisting individuals and legal practitioners in predicting their cases before approaching the court.
Keywords: Deep Learning, Text Mining and Summarization, Collocation,,