The Research Of Warning System For China’s Social Medical Insurance Fund

The social medical insurance fund in China is currently facing a huge surplus as well as a deficit,together with other operational problems,which means that improvements must be made on this system.Thus,we construct a social medical insurance fund warning system,which optimizes fund operation and reduces its cost by giving early warnings.In this paper,we study medical insurance for urban workers.From the medical insurance fund,the hospitals and the economic environment,we extract 16 risk indicators for a warning system.Based on the risk index that reflects the fund’s paying capability,we categorize warnings into deficit warnings and surplus warnings.On this basis,from a new-perspective,we use artificial neural network to build a non-linear relationship between risk indicators of above three aspects and the risk index.The data we use for our empirical research is from Sanming,which is known for obtaining preliminary success in the medical insurance reform.Firstly,we build a Back Propagation(BP)network model,and then,to best fit the data,which are strongly coupling and small in sample sizes,we conduct principal component analysis(PCA)for dimension reduction to building a Principal Component Analysis(PCA)-BP network model.We validate if those models are feasible for modelling and forecasting the social medical insurance fund warning system.Besides,we perform sensitivity analysis for 16 indicators using improved Garson method.Empirical results show that BP network is appropriate for building China’s social medical insurance fund warning system.Compared to BP network model,PCA-BP network model has better data fitting and forecasting accuracy.Moreover,from sensitivity analysis we find indicators that have direct relationships with warnings have higher network connection weights.Lastly,in terms of the introduction and application of the warning system,we come up with several recommendations for sustainable development of the social medical insurance system in China.