고객집단별 보험금에 대한 소지역 추정 = Small area estimation of the insurance benefit for customer segmentations

다국어 초록 (Multilingual Abstract)

Bayesian methods have been focused in recent years for solving small area estimation problems. In this paper, the hierarchical Bayes procedure is implemented via MCMC techniques and compared with the results of One-way, GLM-Normal, and GLM-Gamma cases by analyzing real data of insurance benefit for customer segmentations. After analyzing insurance benefit real data for customer segmentations, we can conclude that the insurance benefit estimator through the small area estimation is more efficient than the estimators by other methods. In addition, we found that the small area estimation gave accurate estimation result for the small number domains.

참고문헌 (Reference)

1 성나영, "소지역 통계 생산을 위한 추정방법" 11 : 111-126, 2000

2 박종태, "소지역 추정법에 관한 비교연구" 12 : 47-55, 2001

3 Geman, S, "Stochastic relaxation, Gibbs distributions and the Bayesian restoration of image" 6 : 73-90, 1984

4 Ghosh, M, "Small area estimation: An Appraisal" 9 : 55-99, 1994

5 김달호, "Small Domain Estimation of the ProportionUsing Survey Weights" 한국데이터정보과학회 18 (18): 1179-1189, 2007

6 김영원, "Small Area Estimation of Unemployment Rate for the Economically Active Population Survey" 한국데이터정보과학회 15 (15): 1-10, 2004

7 Gelfand, A. E, "Sampling-based approaches to calculating marginal densities" 85 : 398-409, 1990