Presented by Kyusuh Park at CSA 2022 International Conference
Mortality Forecasting Using Data Augmentation
Kyusuh Park 1
1 KS Accounting and Actuarial Consulting Co., Ltd.
Abstract. Recent mortality forecasting studies using artificial neural networks (ANNs) have shown improved forecasting performance compared with previous studies, where annual mortality rates were used. The use of annual mortality rates data leads to a problem in that the data are insufficient. Therefore, in this study, mortality rates were forecast by applying the related time series data augmentation methods with an ANN, unlike in existing related studies. The experimental results showed that ANNs with augmented data have improved mortality forecasting performance compared to the case without data augmentation.
Keywords: Data Augmentation, Mortality Forecasting, Artificial Neural Networks, Lee-Carter Model, LSTM, Time Series Data
'Accounting & Actuarial Science' 카테고리의 다른 글
[IFRS17] 보험회사 소급법, 전진법 논란 이해하기 (0) | 2023.07.25 |
---|---|
[IFRS17] CSM 이해를 위한 기본모형 이해 (0) | 2023.07.23 |
[인공지능] 데이터 증강을 이용한 인공신경망 기반 사망률 예측에 관한 연구 (2022) (0) | 2023.06.30 |
[통계] 일반적으로 사용되는 통계테스트 선택할 때 유용 (0) | 2023.06.27 |
[리스크관리] 대리인문제를 통한 리스크관리 재고찰 (0) | 2023.06.23 |