National Basic Medical Insurance Fraud Intelligent Monitoring Research

With the continuous development of China's medical insurance system,the medical insurance system has played an important role in the majority of the people's visits.However,due to imperfect relevant laws and regulations,medical insurance fraud has also become more and more serious,which has affected the healthy and sustainable development of medical insurance in China.Therefore,it is imperative to accurately investigate and punish those who have fraudulent acts in social health insurance funds and increase the strength of anti-fraud measures to ensure the safety of medical insurance funds.This paper has combed the relevant studies of domestic and foreign medical insurance fraud identification,and proposed the methods used in this paper based on the existing research deficiencies.Based on the division of the duration of the statistical cycle,the intelligent monitoring indicators system for basic medical insurance fraud was established from two dimensions of long-term dimension and short-term dimension.Based on the various indicators in the intelligent monitoring system,the unbalanced samples were resampled using the mixed sampling method of clustering and sampling.Lasso regression was used to select variables.Finally,the basic forest insurance was established using the random forest and gradient-elevation decision tree algorithm.The smart monitoring model of fraudulent behavior,the evaluation and optimization of the model based on the recall rate,the correct rate and other indicators,and finally get an intelligent monitoring model that can accurately identify the medical insurance fraud behavior.In theory,this paper constructs a set of comparatively perfect medical insurance intelligent monitoring system from multiple angles,which lays the foundation for the identification of medical insurance fraud in this paper,and can also serve as the data foundation for monitoring the abnormal behavior of medical insurance in the future;in practice,A combination model that can accurately identify fraudulent behaviors of medical insurance is proposed.Compared with the single model,the recognition rate of fraudulent behaviors is improved,and the fraudulent user recognition rate is as high as 93.5%.The research work in this paper provides important reference for anti-fraud work and has certain practical significance for the development of medical social insurance work.