Local Credit Evaluation System Based On Internet Information

In the new era,the Central Committee of the Communist Party of China has been vigorously promoting integrity construction and establishing an honest and trustworthy social environment.Nowadays,most existing credit evaluation research mainly focuses on the financial credit investigation of individuals and enterprises,and a large number of works are done by manual means.It will effectively promote the construction of a social credit system if information technology is used for integrity assessment.Therefore,a local credit evaluation system is designed and developed in this thesis based on Internet information through natural language processing technologies.Detailed work of this thesis can be summarized as follows:First of all,a set of quantitative evaluation indicators for local integrity is designed.With reference to the existing integrity evaluation indicators and official meetings and documents on promoting an honest society,indicators including creditworthiness in government affairs,commercial creditworthiness,social creditworthiness,and judicial credibility are proposed.Then the positive and negative attributes of news in the above-mentioned fields are used as detailed attributes to evaluate the integrity.Meanwhile,the weights on each indicator are established by the analytic hierarchy process.Secondly,a database is designed and built for integrity evaluation.Based on the characteristic analysis of Internet information,Internet news has been collected,extracted,and preprocessed.Furthermore,administrative division dictionaries and place-name recognition algorithms are adopted to identify the place where the news happened,which is taken as a text material for credit evaluation.Thirdly,based on the comparison of multiple classification methods,it is proposed to adopt Light Gradient Boosting Machine(Light GBM)together with Support Vector Machine(SVM)models for classifying whether the obtained text data belongs to integrity news or not.Among them,the positive and negative attributes of integrity news are divided.Moreover,the particle swarm algorithm is applied to optimize Light GBM and SVM classification models,which improves the classification performance.A quantitative assessment of the government integrity is carried out for the localities in the sample set.Finally,the prototype of the local credit evaluation system is designed and developed.The main functions of this system including adding,deleting,modifying,and checking data,secondlevel classification of integrity news,and quantitative evaluation of local creditworthiness.In the end,it is proved that the system has achieved the expected goals by testing the various functions of the system.