2022Afrika StatistikaOpen access

Modelling Skewed and Heavy-tailed Data Using a Normal Weighted Inverse Gaussian Distribution

Calvin B. Maina, Patrick Weke, Carolyne Adhiambo Ogutu, Joseph Makoteku Ottieno

Open full text 0 citations

Abstract

The normal distribution is inadequate in capturing skewed and heavy-tailed behaviour of data taken over short time intervals. In addition the data can be leptokurtic. For this reason a normal weighted inverse Gaussian distribution is proposed as an alternative to the normal inverse Gaussian distribution to handle such data. The mixing distribution used in the normal variance mean mixture is a finite mixture of two special cases of Generalized Inverse Gaussian \((\textit{GIG})\) distribution. The two special cases and the finite mixture are weighted inverse Gaussian distribution. The motivation for this work is that a finite mixture is more flexible than a single/standard distribution. The \(textit{EM}\)-algorithm has been used for parameter estimation.

About this research paper

What this paper is about

The normal distribution is inadequate in capturing skewed and heavy-tailed behaviour of data taken over short time intervals. In addition the data can be leptokurtic. For this reason a normal weighted inverse Gaussian distribution is proposed as an alternative to the normal inverse Gaussian distribution to handle such data. The mixing distribution used in the normal variance mean mixture is a finite mixture of two special cases of Generalized Inverse Gaussian \((\textit{GIG})\) distribution. The two special cases and the finite mixture are weighted inverse Gaussian distribution. The motivation for this work is that a finite mixture is more flexible than a single/standard distribution. The \(textit{EM}\)-algorithm has been used for parameter estimation.

Why it matters

A significance statement is not available in the OpenAlex record.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

The normal distribution is inadequate in capturing skewed and heavy-tailed behaviour of data taken over short time intervals. In addition the data can be leptokurtic. For this reason a normal weighted inverse Gaussian distribution is proposed as an alternative to the normal inverse Gaussian distribution to handle such data. The mixing distribution used in the normal variance mean mixture is a finite mixture of two special cases of Generalized Inverse Gaussian \((\textit{GIG})\) distribution. The two special cases and the finite mixture are weighted inverse Gaussian distribution. The motivation for this work is that a finite mixture is more flexible than a single/standard distribution. The \(textit{EM}\)-algorithm has been used for parameter estimation.

Key concepts: Inverse Gaussian distribution, Kurtosis, Normal-inverse Gaussian distribution, Generalized inverse Gaussian distribution, Normal distribution, Gaussian, Mathematics, Mixture model

Related papers

Back to paper searchBrowse research topicsOriginal source
Modelling Skewed and Heavy-tailed Data Using a Normal Weighted Inverse Gaussian Distribution — Research Paper | ScholarLens