ARTIFICIAL INTELLIGENCE AND LAW: A COMPUTATIONAL SOLUTION CAPABLE OF PREDICTING JUDICIAL DECISIONS

In this paper, we present the result of a computational solution capable of carrying out the prediction of judicial decisions (judicial sentences). We classify plaintiffs of public civil actions of administrative improbity and enforcements of conduct adjustment terms presented by State Prosecution of Parana (in Brazil) from 2011 to 2018, at the State Courts of Parana. The goal of these plaintiffs is to predict judicial decisions (whether favorable or not in relation to the requests made by the State Prosecution's plaintiff). Through natural language processing techniques and machine learning, the results obtained are promising, reaching 78,02% accuracy with the use of the Logistic Regression inductive algorithm, representation based on Bag of Words and using RSLPS stemming.