KNOWLEDGE BASED SYSTEM FOR LAW VIOLATION CLASSIFICATION

• Main Author: Lambarita Nadeak, Leo
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/43107
• Institution: Institut Teknologi Bandung

Legal fiction theory states that every single person in Indonesia is considered to have understood the law, so that justification like “I don’t know about that law” can’t be used as justification in court. But in reality, does everyone really understand the law? Law is a complex entity, dynamic entity, and legal-related events occur frequently. Because of that, society needs an intelligence system that can answer its user by giving appropriate law article that violated in user input. To implement the system, system needs a flexible and scalable knowledge in order to adjust to Indonesian law that already told before. Because of that, the selected knowledge structure to implement the system is Ripple Down Rule (RDR. Apart from the knowledge, system must be able to help user to keep understanding the input needed by system, i.e. by accommodate user input as a text which contains event what the user wants to tell. This text needs to be processed to get actions that happen in the event. The implementation performed by using Natural Language Processing (NLP) module who can process user text, finding words which can be used as action indicator. System that has been implemented is tested to know if system already worked well functionally. Apart from functionality, system also tested from non-functionality aspect by testing its efficiency and effectiveness. System efficiency is tested by calculating system time to do certain function. While the effectiveness of system is tested by comparing a verdict of several case in reality with the conclusions given by the system in same case which is obtained from Indonesian Supreme Court public online repository. Every case document from that repository already contain flow of the violation events and also the verdict of the case.