생명 보험 자료에 대한 생존분석 연구
- Abstract
- This research used the contents of the contracts and policies of xx insurance company (Data from Jan 1, 2011 to Dec 31, 2016) in order to understand the factors leading to a long-term contract which makes up an important part of the life insurance industry in Korea. This research used the Cox's proportional hazards (PH) model and penalized variable selection methods to look for variables that led to a long-term contract. Variables considered include the change of advisor, sex, disease status (anamnesis), risk of job, subscription types, investment patterns of variable insurance, personal versus corporate contracts and driving status, etc.
The most significant factor was the advisor (in case where the recruiter and the collector were different) status, whether it has been changed or not. The second most important factor was the reward type 2 (annuity insurance). Then, the variable insurance investment tendency 4 (safety-pursuing type), and 5 (risk-aversion type) affected the outcome respectively.
This research explores the ways in which the life insurance industry can prolong the contracts, the new market for it, the necessity of the advisors’ long-term commitment, the program consulting, and the improvement of the investment aspects.
- Author(s)
- 양대건
- Issued Date
- 2018
- Awarded Date
- 2018. 8
- Type
- Dissertation
- Publisher
- 부경대학교
- URI
- https://repository.pknu.ac.kr:8443/handle/2021.oak/14653
http://pknu.dcollection.net/common/orgView/200000108993
- Alternative Author(s)
- YANG. DAE GEON
- Affiliation
- 부경대학교 대학원
- Department
- 대학원 통계학과
- Advisor
- 하일도
- Table Of Contents
- 제 1장 서론 1
1.1 연구의 배경 및 목적 1
1.2 연구의 범위 및 구성 4
제 2장 생명보험 자료를 활용한 콕스 비례위험 모형 분석 8
2.1 생존 분석 및 콕스 비례위험 모형의 기본 개념 8
2.2 콕스 비례위험 모형의 형태 8
2.3 콕스 비례위험 변수 선택 9
제 3장 실제 자료 분석 12
3.1 기초 통계분석 12
3.2 카플란 마이어 분석 15
3.3 콕스 비례위험 모형의 분석 18
3.4 콕스 비례위험 모형에서 변수 선택 19
제 4장 결론 및 제언 21
참고문헌 24
부록 : R codes 26
- Degree
- Master
-
Appears in Collections:
- 대학원 > 통계학과
- Authorize & License
-
- Files in This Item:
-
Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.