Design And Implementation Of Recommendation System Of Education Correction Programs For Community Correction Objects

With the continuous incrcase in the number of community prisoners,the risks and challenges for correction targets have become increasingly prominent.The formulation of an education correction program requires community corrections workers to formulate a personalized education correction program that "differs from person to person" based on the idea of inquiry,combined with the background information of the correction object and their performance during correction.At present,the formulation of educational correction programs has problems that the individualization of the educational correction programs is not prominent and the target is poor.The main reasons are as follows:Firstly,there are many background documents and legal documents about the basic information of the correction object,criminal facts,family background,performance during correction,etc.,leading to insufficient preparations for the formulation of education correction programs.Secondly,there are subjective differences when community corrections workers analyze the personality traits,which makes it difficult to guarantee the accuracy and consistency of the education correction plan.Finally,as the human brain has limited data processing capabilities relative to machine learning algorithms,it is unable to efficiently and accurately complete the recommendation logic of education correction programs.In order to solve the above three main problems and recommend targeted and effective education correction measures for prisoners,the main work of this article is as follows:(1)Build a case database.Analyze the data source of the pilot area,establish a target database,design a data migration plan,and perform data preprocessing and attribute construction.Help staff conveniently and quickly query and edit information on prisoners and their education and correction program data.(2)This article uses the method of portrait construction to explore the behavior characteristics of the correction object,the correction appeal,and the focus of the education correction plan.The portraits of prisoners in the community include attitudes toward confession,risk of re-offending,mental health,and social adaptability.The portrait of the education correction program includes the text category,keywords,and topics of the program.(3)This article models the education correction program recommendation problem and designs two recommendation strategies to provide algorithm logic for the education correction program automatic recommendation problem.In order to solve the cold start problem,a collaborative filtering recommendation algorithm based on correction objects can be used.The other is a recommendation algorithm based on FFM-FTRL.Through experimental verification,the optimized model has improved prediction accuracy and training speed.Finally,this paper designs an intelligent recommendation system for education correction schemes,and shows the system prototype.