TY - GEN
T1 - Forecasting mortality rates of elderly in Indonesia using the first generalized Cairns-Blake-Dowd model
AU - Kafi, Rahmat Al
AU - Mardiyati, Sri
AU - Malik, Maulana
N1 - Funding Information:
This research is funded by PITTA 2018 research grant from DRPM Universitas Indonesia (project ID: 2309/UN2.R3.1/HKP.05.00/2018).
Publisher Copyright:
© 2019 Author(s).
PY - 2019/12/4
Y1 - 2019/12/4
N2 - The insurance companies in Indonesia is always aware of the importance of forecasting future mortality rates. If information about mortality rates for several future periods can be obtained in the present time, then the financial planning and policy in determining the amount of premium are expected to be better and more directed. This paper proposed the First Generalized Cairns-Blake-Dowd (CBD) model to forecast the mortality rates of the Indonesian population. The CBD model contains time series parameters. The first stage is to use the Least Square and Newton-Raphson methods to estimate these parameters. The accuracy of the estimation results is verified by the Mean Squared Error (MSE) value. The second stage is to use Holt's linear trend method to forecast the values of estimated parameters aims to calculate the future mortality rates. The level of forecasting accuracy is verified by Mean Absolute Percentage Error (MAPE) value. The MSE and MAPE for each age group are less than 10-4 and 5%, respectively. The forecasted mortality rates for the next periods are presented in the tabular and graphical form. The trend of the projected mortality rate shows a downward trend for each age group.
AB - The insurance companies in Indonesia is always aware of the importance of forecasting future mortality rates. If information about mortality rates for several future periods can be obtained in the present time, then the financial planning and policy in determining the amount of premium are expected to be better and more directed. This paper proposed the First Generalized Cairns-Blake-Dowd (CBD) model to forecast the mortality rates of the Indonesian population. The CBD model contains time series parameters. The first stage is to use the Least Square and Newton-Raphson methods to estimate these parameters. The accuracy of the estimation results is verified by the Mean Squared Error (MSE) value. The second stage is to use Holt's linear trend method to forecast the values of estimated parameters aims to calculate the future mortality rates. The level of forecasting accuracy is verified by Mean Absolute Percentage Error (MAPE) value. The MSE and MAPE for each age group are less than 10-4 and 5%, respectively. The forecasted mortality rates for the next periods are presented in the tabular and graphical form. The trend of the projected mortality rate shows a downward trend for each age group.
UR - http://www.scopus.com/inward/record.url?scp=85076669449&partnerID=8YFLogxK
U2 - 10.1063/1.5136386
DO - 10.1063/1.5136386
M3 - Conference contribution
AN - SCOPUS:85076669449
T3 - AIP Conference Proceedings
BT - Proceedings of the International Conference on Mathematical Sciences and Technology 2018, MathTech 2018
A2 - Yatim, Yazariah Mohd
A2 - Ahmad, Syakila
A2 - Ismail, Mohd Tahir
A2 - Ali, Majid Khan Majahar
A2 - Rahman, Rosmanjawati Abdul
A2 - Sulaiman, Hajar
A2 - Ramli, Norshafira
A2 - Ahmad, Noor Atinah
A2 - Abdullah, Farah Aini
PB - American Institute of Physics Inc.
T2 - 1st International Conference on Mathematical Sciences and Technology 2018: Innovative Technologies for Mathematics and Mathematics for Technological Innovation, MathTech 2018
Y2 - 10 December 2018 through 12 December 2018
ER -