{"title":"使用 ARIMA、ARIMA-GARCH 和 ANN 预测死亡率的多人口奥黑尔模型","authors":"Nurul Syuhada Samsudin, Siti Rohani Mohd Nor","doi":"10.11113/matematika.v39.n3.1496","DOIUrl":null,"url":null,"abstract":"Multi-population mortality model has gained attention from prominent researchers of mortality due to its ability to provide biologically reasonable forecast. Previously, many researchers have proposed several multi-population stochastic mortality models that they considered adequate to produce accurate life expectancy. However, little have been addressed of the variability in full ages and time, which can contribute to an erroneous estimation of life expectancy. Therefore, this study proposed a new multi-population O’Hare with ARIMA, ARIMA-GARCH and ANN in forecasting the mortality rate for male and female in Malaysia, Taiwan, Japan, Hong Kong, Australia, USA, UK, Canada, and Switzerland. Multi-population O’Hare was used as a reference model, whilst ARIMA, ARIMA-GARCH and ANN were incorporated to the reference model to forecast the mortality rates. The adequacy of the proposed model was assessed by using measurement errors which were Mean Absolute Percentage Error (MAPE) and Root Mean Square Error (RMSE). The results showed by multi-population O’Hare with ARIMA-GARCH gave the best forecasting performance for Taiwan, Japan, Australia, USA, UK, Canada, and Switzerland. On the other hand, multi-population O’Hare with ARIMA gave the best forecasting performance for Malaysia, whereas multi-population O’Hare with ANN gave the best forecasting performance for Hong Kong.","PeriodicalId":43733,"journal":{"name":"Matematika","volume":null,"pages":null},"PeriodicalIF":0.3000,"publicationDate":"2023-11-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Multi-Population O’Hare with ARIMA, ARIMA-GARCH and ANN in Forecasting Mortality Rate\",\"authors\":\"Nurul Syuhada Samsudin, Siti Rohani Mohd Nor\",\"doi\":\"10.11113/matematika.v39.n3.1496\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Multi-population mortality model has gained attention from prominent researchers of mortality due to its ability to provide biologically reasonable forecast. Previously, many researchers have proposed several multi-population stochastic mortality models that they considered adequate to produce accurate life expectancy. However, little have been addressed of the variability in full ages and time, which can contribute to an erroneous estimation of life expectancy. Therefore, this study proposed a new multi-population O’Hare with ARIMA, ARIMA-GARCH and ANN in forecasting the mortality rate for male and female in Malaysia, Taiwan, Japan, Hong Kong, Australia, USA, UK, Canada, and Switzerland. Multi-population O’Hare was used as a reference model, whilst ARIMA, ARIMA-GARCH and ANN were incorporated to the reference model to forecast the mortality rates. The adequacy of the proposed model was assessed by using measurement errors which were Mean Absolute Percentage Error (MAPE) and Root Mean Square Error (RMSE). The results showed by multi-population O’Hare with ARIMA-GARCH gave the best forecasting performance for Taiwan, Japan, Australia, USA, UK, Canada, and Switzerland. On the other hand, multi-population O’Hare with ARIMA gave the best forecasting performance for Malaysia, whereas multi-population O’Hare with ANN gave the best forecasting performance for Hong Kong.\",\"PeriodicalId\":43733,\"journal\":{\"name\":\"Matematika\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.3000,\"publicationDate\":\"2023-11-30\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Matematika\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.11113/matematika.v39.n3.1496\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q4\",\"JCRName\":\"MATHEMATICS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Matematika","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.11113/matematika.v39.n3.1496","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"MATHEMATICS","Score":null,"Total":0}
引用次数: 0
摘要
多人口死亡率模型因其能够提供生物学上合理的预测而受到死亡率领域著名研究人员的关注。在此之前,许多研究人员提出了几种多人口随机死亡率模型,他们认为这些模型足以得出准确的预期寿命。然而,这些模型很少涉及全年龄和时间的变异性,而这种变异性可能导致对预期寿命的错误估计。因此,本研究提出了一种新的多人口 O'Hare,用 ARIMA、ARIMA-GARCH 和 ANN 预测马来西亚、台湾、日本、香港、澳大利亚、美国、英国、加拿大和瑞士的男性和女性死亡率。多人口 O'Hare 被用作参考模型,而 ARIMA、ARIMA-GARCH 和 ANN 被纳入参考模型以预测死亡率。利用测量误差(平均绝对百分比误差 (MAPE) 和均方根误差 (RMSE))评估了拟议模型的适当性。结果显示,在台湾、日本、澳大利亚、美国、英国、加拿大和瑞士,多人口 O'Hare 与 ARIMA-GARCH 的预测效果最佳。另一方面,使用 ARIMA 的多人口 O'Hare 对马来西亚的预测效果最好,而使用 ANN 的多人口 O'Hare 对香港的预测效果最好。
Multi-Population O’Hare with ARIMA, ARIMA-GARCH and ANN in Forecasting Mortality Rate
Multi-population mortality model has gained attention from prominent researchers of mortality due to its ability to provide biologically reasonable forecast. Previously, many researchers have proposed several multi-population stochastic mortality models that they considered adequate to produce accurate life expectancy. However, little have been addressed of the variability in full ages and time, which can contribute to an erroneous estimation of life expectancy. Therefore, this study proposed a new multi-population O’Hare with ARIMA, ARIMA-GARCH and ANN in forecasting the mortality rate for male and female in Malaysia, Taiwan, Japan, Hong Kong, Australia, USA, UK, Canada, and Switzerland. Multi-population O’Hare was used as a reference model, whilst ARIMA, ARIMA-GARCH and ANN were incorporated to the reference model to forecast the mortality rates. The adequacy of the proposed model was assessed by using measurement errors which were Mean Absolute Percentage Error (MAPE) and Root Mean Square Error (RMSE). The results showed by multi-population O’Hare with ARIMA-GARCH gave the best forecasting performance for Taiwan, Japan, Australia, USA, UK, Canada, and Switzerland. On the other hand, multi-population O’Hare with ARIMA gave the best forecasting performance for Malaysia, whereas multi-population O’Hare with ANN gave the best forecasting performance for Hong Kong.