Research On Domain-oriented Prediction Analysis Methods And Its Applications

Different from other insurances in the social insurance system, sustainability is one of thecore issues to China’s basic endowment insurance system. It is important to the stability anddevelopment of social economic. Currently, the researches on the sustainability of China’s basicendowment insurance are mainly in using of actuarial methods, besides, most of the researchesuse statistical data leads to the proposed models are often very lopsided and the results are opento question because of lacking the validation of the real data.For the above, in this thesis we give out the sustainability analysis models under analyzingthe policy and real data of the basic endowment insurance from the micro-view and macro-view.When analyze the sustainability from the macro-view, it needs mass of macro-analysis-data,however, it is found that social insurance information systems are not uniform in practice forany reason that cause the existing system architecture cannot meet the requirement ofmacro-analysis, so it needs to build a dataware house of basic endowment insurance from themacro-analysis-view. Although, using analysis models with real data and certain assumptionscan grasp better on the sustainability trends of basic endowment insurance fund, however,because there exist assumptions in the analysis process lead to the results not precise enough.So, we work out an effective prediction model to predict the indicators in the actuarial modeland eliminate the assumptions so that the analysis is all under real data. At last, we design aprediction system of basic endowment insurance.Whether the prediction model is good or not can directly influence the quality ofmacro-analysis. There is a prediction model formed in this paper on research of variousquantitative prediction methods and based on the theories of artificial neural networkintegration and selective integration, which is used into the practice of the prediction. Throughthe comparison, it can be seen that it has a better ability on fitting the real data of basicendowment insurance.