Research And Implementation Of Classification Model Of Case Settlement Methods Based On Ensemble Learning

In the context of the rapid development of modern science and technology,the combination of information construction and the standardization of the implementation work of judicial organs are inevitable to avoid cloud computing and artificial intelligence.Strengthening the top-level design planning of the court,promoting the development of the court's execution toward the direction of intelligence,integration and practicality,and scientifically using modern technological means to solve the traditional execution problems,have become the top priority of the court's informatization construction.The main work of this paper is to study the use of artificial intelligence technology to predict the conclusion method of execution cases,and thus help solve problems in enforcement.On the one hand,this paper can help the applicant to form the best execution strategy,so as to save the execution cost.On the other hand,it can help judges make use of big data and artificial intelligence technology to ensure judicial fairness and justice.The employment of cloud desktop technology can solve the problem of low efficiency of traditional PC operation and maintain the data security.The application of cloud computing and big data technology in the informatization of court execution is a product of the combination of deepening the reform of the judicial system and the application of modern science and technology,which can significantly improve the quality,efficiency and accuracy of execution work.In this paper,this article first introduces the theory of the used technology and the corresponding advantages and disadvantages,then designs a set of data cleaning and coding methods.And uses the association rule algorithm to discover the rules related to the case settlement methods.In addition,in terms of the importance of variables,the information gain method is adopted to explore the influence of different characteristics on the way of case conclusion,and uses SMOTE to solve imbalance data,finally employs the H2O-based framework and distributed Hadoop platform to explore the performance of the random forest and automated machine learning algorithm and neural network.Secondly,based on the data obtained in Chapter 2,Chapter 3 employs the Stacking method composed of integrated K-nearest neighbor algorithm,Naive Bayes algorithm,decision tree algorithm and random forest algorithm to classify the closing method,after comparing with the voting classifier method,results show that the Stacking algorithm classification is more accurate.Finally,the virtual nodes system management module is designed based on the VMware v Sphere cloud desktop,and two security authentication methods are designed to ensure the security.