Researchment On Risk Control Of Internet Consumer Finance Based On Character Analysis

At the end of 2013,the Third Plenary Session of the 18 th CPC Central Committee established the development of inclusive finance as a national strategy.At the end of 2015,the “Promoting Inclusive Financial Development Plan” was released,and Internet consumer finance,as a pioneer of inclusive finance,developed rapidly.Among them,licensed consumer finance companies,p2 p lending companies,and e-commerce platforms accounted for e-commerce finance.More than half of the consumer finance sector.On the one hand,Internet consumer finance has broken the barriers of traditional banks,enabling borrowers to directly achieve financing online.This innovative financial model has brought new opportunities to China's financial reform,and has injected new opportunities for China's economic transformation and consumption upgrading.Vitality.On the other hand,the negative news of Internet consumer finance has appeared frequently,and its risk control ability has been questioned.Therefore,how to do a good job of risk control and form a healthy profit model will be the primary problem facing Internet consumer finance.Analysis of the development status of China's Internet consumer finance,Internet consumer finance products are mainly divided into three business models: first,information capital service intermediary: second,the traditional consumer finance Internet;third,based on e-commerce platform to carry out financial business.Under the trend of rapid development of Internet consumer finance,there are mainly three types of risks,credit risk,technical risk and regulatory risk.Among them,fraud risk is the main form of credit risk,and it is the main indicator of Internet consumer finance risk control ability.Therefore,the improvement of Internet consumer finance anti-fraud ability cannot be delayed.Based on the research results at home and abroad,combined with the knowledge of disciplines such as statistics and machine learning,and based on big data,this paper proposes a set of anti-fraud models for more reasonable Internet consumer finance.The model is mainly composed of statistical analysis,model verification,and manual verification.Statistical verification uses statistical methods to collect,organize and analyze raw data.Model verification uses machine learning algorithms to establish an anti-fraud model to verify fraud.Sexual users;manual verification is a supplement to the first two modules,again reviewing users who may be fraudulent.Based on the model framework of the anti-fraud system,this paper collects the actual data of the industry of Internet consumer finance,and uses it to model and analyze.The raw data set contains the user's consumption information,social information,and credit information(exception type and normal type).Based on statistical analysis of big data,it is found that the user's consumption and social interaction are positively related to the user's credit type.The user's social relationship has a significant impact on the user's credit rating.The neural network model,the support vector machine model and the random forest model are respectively extracted from the training set data,and the test set data is used for verification.The model evaluation shows that the neural network has the best prediction effect.Finally,it summarizes the research results of the Internet consumer finance anti-fraud system,expounds the shortcomings in the research,and gives reasonable opinions on the construction of the Internet consumer finance anti-fraud system.