TY - GEN

T1 - Discrete Lindley distribution

AU - Diandarma, R.

AU - Nurrohmah, S.

AU - Fithriani, I.

N1 - Publisher Copyright:
© 2020 Author(s).

PY - 2020/6/1

Y1 - 2020/6/1

N2 - Overdispersion often being a problem in modeling count data because the Poisson distribution that is often used to modeling count data cannot conquer the overdispersion data. Several distributions have been introduced to be used as an alternative to the Poisson distribution on conquering dispersion in data. However, that alternative distribution has higher complexity than Poisson distribution in the number of parameters used. Therefore, a new distribution with similar distribution to Poisson is offered, which is Lindley distribution. Lindley distribution is a continuous distribution, then it cannot be used to modeling count data. Hence, discretization on Lindley distribution should be done using a method that maintains the survival function of Lindley distribution. Result distribution from discretization on Lindley distribution has one parameter and can be used to modeling overdispersion data so that distribution is appropriate to be used as an alternative to Poisson distribution in modeling overdispersed count data. The result distribution of Lindley distribution discretization is commonly called Discrete Lindley distribution. In this paper, characteristics of Discrete Lindley distribution that are obtained are unimodal, right skew, high fluidity and overdispersion. Based on numerical simulation, another characeristic of parameter is also obtained from Discrete Lindley distribution that has a large bias and MSE when parameter value around e-1.

AB - Overdispersion often being a problem in modeling count data because the Poisson distribution that is often used to modeling count data cannot conquer the overdispersion data. Several distributions have been introduced to be used as an alternative to the Poisson distribution on conquering dispersion in data. However, that alternative distribution has higher complexity than Poisson distribution in the number of parameters used. Therefore, a new distribution with similar distribution to Poisson is offered, which is Lindley distribution. Lindley distribution is a continuous distribution, then it cannot be used to modeling count data. Hence, discretization on Lindley distribution should be done using a method that maintains the survival function of Lindley distribution. Result distribution from discretization on Lindley distribution has one parameter and can be used to modeling overdispersion data so that distribution is appropriate to be used as an alternative to Poisson distribution in modeling overdispersed count data. The result distribution of Lindley distribution discretization is commonly called Discrete Lindley distribution. In this paper, characteristics of Discrete Lindley distribution that are obtained are unimodal, right skew, high fluidity and overdispersion. Based on numerical simulation, another characeristic of parameter is also obtained from Discrete Lindley distribution that has a large bias and MSE when parameter value around e-1.

KW - Count data

KW - Discrete Lindley distribution

KW - Lindley distribution

KW - overdispersion

KW - Poisson distribution

UR - http://www.scopus.com/inward/record.url?scp=85086665647&partnerID=8YFLogxK

U2 - 10.1063/5.0007950

DO - 10.1063/5.0007950

M3 - Conference contribution

AN - SCOPUS:85086665647

T3 - AIP Conference Proceedings

BT - Proceedings of the 5th International Symposium on Current Progress in Mathematics and Sciences, ISCPMS 2019

A2 - Mart, Terry

A2 - Triyono, Djoko

A2 - Ivandini, Tribidasari Anggraningrum

PB - American Institute of Physics Inc.

T2 - 5th International Symposium on Current Progress in Mathematics and Sciences, ISCPMS 2019

Y2 - 9 July 2019 through 10 July 2019

ER -