基于Esscher損失函數(shù)和最大熵方法的遞推信度模型.pdf

1、碩士研究生學(xué)位論文新新新疆疆疆大大大學(xué)學(xué)學(xué)論文題目(中文)：基基基于于于Esscher損損損失失失函函函數(shù)數(shù)數(shù)和和和最最最大大大熵熵熵方方方法法法的的的遞遞遞推推推信信信度度度模模模型型型論文題目(英文)：TheRecursiveCredibilityModelsBasedonEsscherLossFunctionMaximumEntropyMethod研究生姓名：張興紅學(xué)科專業(yè)：數(shù)學(xué)學(xué)位類別：理學(xué)碩士研究方向：概率論與數(shù)理統(tǒng)計導(dǎo)師姓名

2、職稱：吳黎軍教授論文答辯日期年月日學(xué)位授予日期年月日AbstractAbstractCredibilitytheyisdevotedtocalculatethecredibilityestimateofpremiumofcrespondingrisksbasedonthethehisticalexperienceofasinglepolicyholdertheentiregroupofpolicyholdersIngeneralthem

3、ostcredibilitypremiumistheweightedsumofaverageoftheindividualpremiumcollectivepremiumtheweightedfactiscalledcredibilityfact.Thismethodhasbeenwidelyappliedtothevariousareasofinsurancepractice.ontheonehweestablishrecursive

4、credibilitymodelundermaximumentropymethodwhichaimstominimumthemeansquarederrmaximumtheentropy.wealsodeducetheexpressionfcredibilitypremiumcredibilityfact.Atthesametimeweshowspecialcasesofcredibilitymodel.onehavelargerwei

5、ghtinrecentyearsanotherhavesmallerweightinrecentyearswhichgeneralizestheclassicalcredibilitythey.ontheotherhthecredibilitypremiumisonthebasisofpurepremiumwhichcannotbeusedintheinsurancepracticetherwiseinsurancecompaniesw

6、illfacebankruptcy.OneofthemethodtosolvetheproblemistochangethelossfunctioninthepaperweaimstoestablishrecursivecredibilitymodelunderEsscherLossfunctiondeducerecursivecredibilityundergeneralizedEsscherLossfunction.finallyw