A finite mixture of multiple discrete distributions for modelling heaped count data

A new modelling approach, based on finite mixtures of multiple discrete distributions of different multiplicities, is proposed to fit data with a lot of periodic spikes in certain values. An EM algorithm is provided in order to ensure the models' ease-of-fit and then a simulation study is prese...

Descripción completa

Detalles Bibliográficos
Autores: Bermúdez, Lluís, Karlis, Dimitris, Santolino, Miguel
Tipo de recurso: artículo
Estado:Versión aceptada para publicación
Fecha de publicación:2017
País:España
Institución:Universidad de Barcelona
Repositorio:Dipòsit Digital de la UB
OAI Identifier:oai:diposit.ub.edu:2445/108966
Acceso en línea:https://hdl.handle.net/2445/108966
Access Level:acceso abierto
Palabra clave:Anàlisi de regressió
Assegurances d'accidents
Variables (Matemàtica)
Regression analysis
Accident insurance
Variables (Mathematics)
Descripción
Sumario:A new modelling approach, based on finite mixtures of multiple discrete distributions of different multiplicities, is proposed to fit data with a lot of periodic spikes in certain values. An EM algorithm is provided in order to ensure the models' ease-of-fit and then a simulation study is presented to show its efficiency. A numerical application with a real data set involving the length, measured in days, of inability to work after an accident occurs is treated. The main finding is that the model provides a very good fit when working week, calendar week and month multiplicities are taken into account.