TY - JOUR

T1 - MODELLING THE NUMBER OF HIV AND AIDS CASES IN EAST JAVA USING BIRESPONSE MULTIPREDICTOR NEGATIVE BINOMIAL REGRESSION BASED ON LOCAL LINEAR ESTIMATOR

AU - Tohari, Amin

AU - Nur, Chamidah

AU - Fatmawati,

AU - Budi, Lestari

N1 - Funding Information:
This research was funded through the Doctoral Dissertation Research Grant with a contract No. 823/UN3.14/PT/2020. Therefore, the authors thank to the Directorate of Research and Public Service, the Ministry of Education and Culture of the Republic of Indonesia.
Publisher Copyright:
© 2021 the author(s).

PY - 2021/1/1

Y1 - 2021/1/1

N2 - A virus which attacks the CD4 lymphocytes of the immune system is called human immunodeficiency virus (HIV). If HIV is not addressed then the disease will develop into acquired immunodeficiency syndrome (AIDS). It means that HIV is the cause behind the AIDS infection. Indonesia has been classified into countries where HIV infection rates are high enough since 2006. East Java province was one of six other provinces in Indonesia entering endemic regions other than Jakarta, Papua, West Java, Riau and Bali. The our interest cases namely the number of HIV and AIDS cases in East Java, are two things that correlate with each other. They are categorized into discrete variables. Therefore, this study aims to model the our interest cases with both drug users and percentage of contraceptive users by using local linear Biresponse Multipredictor Negative Binomial (BMNB) regression model approach. By using maximum likelihood cross validation (MLCV) method, we obtain optimal bandwidth of first predictor variable (drug users) and second predictor variable (percentage of contraceptive users), i.e., 30 and 2.5, respectively. We get deviance values of 0.473 for local linear BMNB regression model approach and of 4.4822 for parametric regression model approach. It means that for modeling the our interest cases, the use of local linear BMNB regression model approach is better than the use of parametric regression model approach.

AB - A virus which attacks the CD4 lymphocytes of the immune system is called human immunodeficiency virus (HIV). If HIV is not addressed then the disease will develop into acquired immunodeficiency syndrome (AIDS). It means that HIV is the cause behind the AIDS infection. Indonesia has been classified into countries where HIV infection rates are high enough since 2006. East Java province was one of six other provinces in Indonesia entering endemic regions other than Jakarta, Papua, West Java, Riau and Bali. The our interest cases namely the number of HIV and AIDS cases in East Java, are two things that correlate with each other. They are categorized into discrete variables. Therefore, this study aims to model the our interest cases with both drug users and percentage of contraceptive users by using local linear Biresponse Multipredictor Negative Binomial (BMNB) regression model approach. By using maximum likelihood cross validation (MLCV) method, we obtain optimal bandwidth of first predictor variable (drug users) and second predictor variable (percentage of contraceptive users), i.e., 30 and 2.5, respectively. We get deviance values of 0.473 for local linear BMNB regression model approach and of 4.4822 for parametric regression model approach. It means that for modeling the our interest cases, the use of local linear BMNB regression model approach is better than the use of parametric regression model approach.

KW - HIV and AIDS

KW - Local linear BMNB regression model

KW - Percentage of drug and contraceptive users

UR - http://www.scopus.com/inward/record.url?scp=85118200852&partnerID=8YFLogxK

U2 - 10.28919/cmbn/5652

DO - 10.28919/cmbn/5652

M3 - Article

AN - SCOPUS:85118200852

SN - 2052-2541

VL - 2021

JO - Communications in Mathematical Biology and Neuroscience

JF - Communications in Mathematical Biology and Neuroscience

M1 - 73

ER -