TY - GEN
T1 - Hierarchical Bayesian Modelling on Predicting East Java Province Population
AU - Chee, Darren Kang Wan
AU - Iriawan, Nur
AU - Widhianingsih, Tintrim Dwi Ary
N1 - Publisher Copyright:
© 2024 IEEE.
PY - 2024
Y1 - 2024
N2 - The human population is constantly increasing every second. With this constantly increasing happening, in the long term it will cause overpopulation, and it will lead to more problems to come. Therefore, the anticipation of the increase or decrease in the population will be needed. This paper proposes using the Hierarchical Bayesian Mixture (HBM) method to predict the population of East Java Province. The main contribution of this work is to build a model that can be used to predict population size as discrete data as a response model based on predictors in an HBM models - compared to a Hierarchical Bayesian non-Mixture (HBNM) model. The HBM model is structured as a form of Poisson Mixture Regression according to the population pattern in each district/city as a model with its covariates at the first level. The variability of all parameters of this first model will be explained by the hyper-parameter distribution as the prior model at the second level. The result indicates that the HBM of the Poisson Mixture Regression model is the better model compared to HBNM of the Poisson Mixture Regression with Deviance Information Criterion of 904,072 and 2,733,250 respectively. This modelling method can be useful for government or policymakers as the basis for anticipating the future growth of the number of populations in East Java Province.
AB - The human population is constantly increasing every second. With this constantly increasing happening, in the long term it will cause overpopulation, and it will lead to more problems to come. Therefore, the anticipation of the increase or decrease in the population will be needed. This paper proposes using the Hierarchical Bayesian Mixture (HBM) method to predict the population of East Java Province. The main contribution of this work is to build a model that can be used to predict population size as discrete data as a response model based on predictors in an HBM models - compared to a Hierarchical Bayesian non-Mixture (HBNM) model. The HBM model is structured as a form of Poisson Mixture Regression according to the population pattern in each district/city as a model with its covariates at the first level. The variability of all parameters of this first model will be explained by the hyper-parameter distribution as the prior model at the second level. The result indicates that the HBM of the Poisson Mixture Regression model is the better model compared to HBNM of the Poisson Mixture Regression with Deviance Information Criterion of 904,072 and 2,733,250 respectively. This modelling method can be useful for government or policymakers as the basis for anticipating the future growth of the number of populations in East Java Province.
KW - hierarchical bayesian mixture
KW - mixture poisson regression
KW - population prediction
UR - https://www.scopus.com/pages/publications/105015682399
U2 - 10.1109/INCITEST64888.2024.11121518
DO - 10.1109/INCITEST64888.2024.11121518
M3 - Conference contribution
AN - SCOPUS:105015682399
T3 - INCITEST 2024 - Proceedings of the 7th International Conference on Informatics, Engineering, Sciences and Technology
BT - INCITEST 2024 - Proceedings of the 7th International Conference on Informatics, Engineering, Sciences and Technology
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 7th International Conference on Informatics Engineering, Science and Technology, INCITEST 2024
Y2 - 24 October 2024
ER -