View : 801 Download: 0

Full metadata record

DC Field Value Language
dc.contributor.advisor안재윤-
dc.contributor.author이영주-
dc.creator이영주-
dc.date.accessioned2020-02-03T16:31:17Z-
dc.date.available2020-02-03T16:31:17Z-
dc.date.issued2020-
dc.identifier.otherOAK-000000163999-
dc.identifier.urihttp://dcollection.ewha.ac.kr/common/orgView/000000163999en_US
dc.identifier.urihttps://dspace.ewha.ac.kr/handle/2015.oak/252736-
dc.description.abstract고전적인 Collective Risk Model는 청구 빈도 간 종속성은 허용하나, 청구 빈도와 청구 심도 사이의 독립성을 가정한다. 따라서 청구 총액을 예측하려면 청구 빈도 내역에 기반한 청구 빈도 예측만으로 충분하다. 즉, 청구 빈도의 예측값에 청구 심도 평균을 곱해줌으로써 청구 총액의 예측값을 얻을 수 있다. 하지만 최근 연구들은 Collective Risk Model에서 청구 빈도와 심도 사이의 종속성뿐만 아니라 청구 심도 간 종속성도 발견하였고, 이는 청구 총액 예측에 청구 빈도 내역과 더불어 청구 심도 내역을 사용하도록 야기했다. 본 연구에서는 청구 심도 내역을 사용하는 것이 항상 청구 총액 예측력을 개선하는 것은 아니기 때문에 주의해야 한다고 주장한다. 특히 청구 심도 간 종속성이 강하지 않은 경우, 이는 보험 업계에서 보편적인 경우인데,에 청구 심도 내역의 추가적인 사용은 청구 빈도 내역만 사용하였을 때보다 청구 총액 예측력을 떨어뜨릴 수 있다. 이를 뒷받침하기 위해 수치적 연구가 동반된다.;In the classical collective risk model, frequency and severity are assumed to be independent while allowing dependence among frequencies. Hence, for the prediction of the aggregate severity, it is enough to predict the frequency based on the frequency history, then the prediction of the aggregate severity is obtained by multiplying the predicted value of frequency and the mean of severity. Recent literatures found the dependence among severities as well as the dependence between frequency and severity in the collective risk model, which leads to use of the severity history in addition to the frequency history for the prediction of the aggregate severity. In this study, we claim that use of the severity history should be cautious since it does not always lead to improvement in the prediction of the aggregate severity. Especially, when the dependence among severities are not strong, which is usual case in most of insurance setting, use of additional severity history can result in poorer prediction performance than using the frequency history only. Numerical study was accompanied to support our claim.-
dc.description.tableofcontentsI. Introduction 1 II. Model Description 3 A. Statistical Models 3 1. Frequency Model in classical BMS 5 2. Copula-Based Random Effect Model in BMS 6 B. Bhlmann Credibility Theory 7 III. Bhlmann Prediction with Frequency Random Effect Model 10 A. Bhlmann Prediction based on the frequency history 11 B. Bhlmann Prediction based on the aggregate severity history 12 IV. Bhlmann Prediction with Frequency and Severity Random Effects Model 13 A. Bhlmann Prediction based on the frequency history 14 B. Bhlmann Prediction based on the aggregate severity history 16 V. Structure of Mean Square Error 18 A. MSE under Frequency Random Effect Model 18 1. Based on the frequency history 18 2. Based on the aggregate severity history 18 B. MSE under Frequency and Severity Random Effects Model 18 1. Based on the frequency history 18 2. Based on the aggregate severity history 20 VI. Numerical Study 22 VII. Conclusion 27 References 28 Abstract(in Korean) 30-
dc.formatapplication/pdf-
dc.format.extent678562 bytes-
dc.languageeng-
dc.publisher이화여자대학교 대학원-
dc.subject.ddc500-
dc.titleChoice between Frequency History and Aggregate Severity History in Collective Risk Model-
dc.typeMaster's Thesis-
dc.title.subtitleTheoretical and Practical Considerations-
dc.creator.othernameYoungJu Lee-
dc.format.pageiii, 30 p.-
dc.identifier.thesisdegreeMaster-
dc.identifier.major대학원 통계학과-
dc.date.awarded2020. 2-
Appears in Collections:
일반대학원 > 통계학과 > Theses_Master
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

BROWSE