Dissertation > Excellent graduate degree dissertation topics show

The Method of Reserve-estimating Based on Bayesian General Linear Model

Author: WangZuo
Tutor: LiuLePing
School: Tianjin University of Finance and Economics
Course: Statistics
Keywords: Outstanding Claim Reserve General Linear Model Bayesian Method Over-dispersed Poisson model
CLC: F840
Type: Master's thesis
Year: 2009
Downloads: 198
Quote: 0
Read: Download Dissertation

Abstract


For the insurance companies which conduct the non-life products, the loss reserves are their biggest liabilities. And the outstanding claim reserves are the most important part of the reserve. So it is essential to choose a suitable method to estimate the outstanding claim reserve.Now some deterministic methods, especially the chain-ladder, attract much popularity to the insurance companies. The deterministic methods are simple to implement, but they only give estimate results and could not examine the estimate so that it could reach the require of the dynamic financial analysis for the insurance companies. So it is a progress to use a stochastic method for reserve instead of the deterministic method. In order to make use of the expert opinion to increase the precision of the prediction, Bayesian theory and method is used for the reserve.In this paper, writer first introduces the General Linear Model (GLM). Next, some theories and methods about Bayesian statistical inference are illustrated, and then stochastic models for estimating outstanding claims reserving are introduced. On the basis of these theories, writer explains how to establish a Bayesian General Linear model which incorporates some prior information to estimate outstanding Claim Reserve. Finally by comparing the result of the Bayesian General Linear model with the results from the chain-ladder and B-F methods, writer has the conclusion that the result from the Over-dispersed Poisson Model with improper prior distribution is similar the traditional method, and the result from the Over-dispersed Poisson Model with strong prior distribution is different from the B-F method. Furthermore, the result from the Negative Binomial model with strong prior distribution is very close to the B-F method, and the result from the Negative Binomial model with weak prior distribution is between the B-F method and the chain-ladder method.

Related Dissertations

  1. Bayesian Methods for Time Series Outliers Detection and Applications in GPS Data Processing,O212.8
  2. The Application of Bayesian Method in Actuarial,F224;F840
  3. The Censored Regression Model under Complex Survey and Its Application in Medical Study,R195
  4. Independent Component Analysis and Its Application in fMRI,R318
  5. Clinical and Genetic Epidemiological Study on Efficacy of Lowering Cholesterol and Toxicity Response to Simvastatin Treatment,R589.2
  6. Characterization of Admissibility for Linear Estimator and Its Superiority in General Linear Model,O212.1
  7. Research on Precision Assessment and Reliability Growth Evaluation Based on Bayesian Inference on Dynamic Population,TJ413.3
  8. The Research of the Risk Management of Venture Capital Companies in China,F832.48
  9. Multi-source information fusion method based on Bayes Small sample theory,TP202
  10. The Study about Methods of VaR Model,O213
  11. Studying on Prior Distribution in Bayesian Learning,TP181
  12. The Determination of Prior Distribution and Bayesian Reliability Evaluation in the Case of Dynamic Collectivity,E920
  13. Research of Reliability Test of Spacecraft Relay Based on Bayesian Method,V242
  14. Continuous Time of Bayesian Model and Empirical Analysis,F830
  15. Research on Bayesian Estimation Methods of the Nonparametric Regressive Model and Its Application,O212.8
  16. Information Retrieval Using Categorization Structures,TP391.3
  17. Application of Kalman Filter in Non-life Outstanding Claims Reserving,F840
  18. Study on Tourism Spot Online Attraction Video Active Service System,TP391.41
  19. China 's consumer metering research,F126
  20. Face Recognition Based on Bayesian Method,TP391.41

CLC: > Economic > Fiscal, monetary > Insurance > Insurance Theory
© 2012 www.DissertationTopic.Net  Mobile