Risk early warning model of complaints and petitions based on random forest algorithm

• 1、School of Big Data and Software, Chongqing University, Chongqing 401331
• 2、Artificial Intelligence Academy of Law, Southwest University of Political Science & Law, Chongqing 401120
• 3、National Academy of Development and Strategy, Remin University of Chna, Beijing 100872

Abstract：Improper handling of complaints and petitions may cause social conflicts and endanger public safety. It is important to maintain social stability to provide risk early warning. It is of great significance to establish a reliable early warning mechanism of complaints and petitions. Data mining techniques were applied to analyze the risk prevention strategies of complaints and petitions, and key information was extracted from the Risk early warning model of complaints and petitions based on random forest algorithmcollected tens of thousands of judgement documents in five provinces and municipalities, then sorted out different dimensions of data as features, then invited legal experts to mark the risk of complaints and petitions based on existing documents. Then a risk early warning model was constructed by training normalized data through random forest algorithm, which was for risk warning of newly generated judgement document data. The experimental verification shows that the model\'s risk prediction accuracy rate reaches 88.10%, which can be effectively applied to the early warning mechanism of the complaints and petitions risk. It as well as will play a positive role in promoting the creation of a smart court.

Keywords： complaints and petitions risk early warning random forest judgement document