Joint Cognition of Both Human and Machine for Predicting Criminal Punishment in Judicial System
Amit Kumar Das, Aziza Ashrafi, Muktadir Ahmmad · 2019 IEEE 4th International Conference on Computer and Communication Systems (ICCCS) · 2019
Thousands of research have been taking place to develop advanced Artificial Intelligence System which can't only perform faster but also predict better than human. But a human has some qualities which can never be gained by a machine like creativity, empathy, sensing, and critical thinking. By aggregating the best sides of both, a novel paradigm for the judicial system can be anticipated. For this purpose, we prepare a dataset both from an online survey (n=103) and interviews conducted in Bangladesh on cases related to `Women and Children Repression Prevention Act, 2000'. We apply several machine learning algorithms to make a machine that can predict punishment like a judge and calculate both the test accuracy and the predictive power of the models to observe which algorithm performs better and stable than the others. Even human can guide machine for judging a delinquent.