A Novel Method of Indian Penal Code Section Classification for Offence Incident Report

Judiciary system a sector of world in which lot of text data (like First information report (FIR), investigation report, trail etc.) generated day to day. Some other document like rule book, legal article also has a lot of text information. Similarity of these documents helps us to take decision in some way. In India, Indian Penal Code is main legal judgment article for defining offence. Any information can be used well only when that information is presented properly and when it comes to computer program, it becomes very important to have the information in a specific structured format. The development of a corpus for the Indian Penal Code (IPC) document is necessary for text similarity between description or keywords of IPC section and offence report related to the judiciary system of India. IPC document is already available as unstructured text in hard and soft copy. This document is used manually by human for reference in judiciary related tasks, but for computerized work it is required in structured and specific format. In this paper, we provide an idea of development of corpus for IPC document and find novel method for classification of IPC Section by comparing different type of word embedding technique for text similarity of two documents and introduce leveled searching method for better result.