Introducing COZIGAM: An R package for unconstrained and constrained zero-inflated generalized additive model analysis

Hai Liu, Kung Sik Chan

Research output: Contribution to journalArticle

24 Citations (Scopus)

Abstract

Zero-inflation problem is very common in ecological studies as well as other areas. Nonparametric regression with zero-inflated data may be studied via the zero-inflated generalized additive model (ZIGAM), which assumes that the zero-inflated responses come from a probabilistic mixture of zero and a regular component whose distribution belongs to the 1-parameter exponential family. With the further assumption that the probability of non-zero-inflation is some monotonic function of the mean of the regular component, we propose the constrained zero-inflated generalized additive model (COZIGAM) for analyzing zero-inflated data. When the hypothesized constraint obtains, the new approach provides a unified framework for modeling zero-inflated data, which is more parsimonious and efficient than the unconstrained ZIGAM. We have developed an R package COZIGAM which contains functions that implement an iterative algorithm for fitting ZIGAMs and COZIGAMs to zero-inflated data based on the penalized likelihood approach. Other functions included in the package are useful for model prediction and model selection. We demonstrate the use of the COZIGAM package via some simulation studies and a real application.

Original languageEnglish
Pages (from-to)1-26
Number of pages26
JournalJournal of Statistical Software
Volume35
Issue number11
StatePublished - Jul 2010

Fingerprint

Generalized Additive Models
Model Analysis
Zero
Generalized additive models
Zero-inflation
Monotonic Function
Penalized Likelihood
Exponential Family
Nonparametric Regression
Model Selection
Inflation
Prediction Model
Iterative Algorithm

Keywords

  • EM algorithm
  • Model selection
  • Penalized likelihood
  • Proportionality constraints

ASJC Scopus subject areas

  • Software
  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Introducing COZIGAM : An R package for unconstrained and constrained zero-inflated generalized additive model analysis. / Liu, Hai; Chan, Kung Sik.

In: Journal of Statistical Software, Vol. 35, No. 11, 07.2010, p. 1-26.

Research output: Contribution to journalArticle

@article{279408a230404ea3ac4145cc7d94f3cc,
title = "Introducing COZIGAM: An R package for unconstrained and constrained zero-inflated generalized additive model analysis",
abstract = "Zero-inflation problem is very common in ecological studies as well as other areas. Nonparametric regression with zero-inflated data may be studied via the zero-inflated generalized additive model (ZIGAM), which assumes that the zero-inflated responses come from a probabilistic mixture of zero and a regular component whose distribution belongs to the 1-parameter exponential family. With the further assumption that the probability of non-zero-inflation is some monotonic function of the mean of the regular component, we propose the constrained zero-inflated generalized additive model (COZIGAM) for analyzing zero-inflated data. When the hypothesized constraint obtains, the new approach provides a unified framework for modeling zero-inflated data, which is more parsimonious and efficient than the unconstrained ZIGAM. We have developed an R package COZIGAM which contains functions that implement an iterative algorithm for fitting ZIGAMs and COZIGAMs to zero-inflated data based on the penalized likelihood approach. Other functions included in the package are useful for model prediction and model selection. We demonstrate the use of the COZIGAM package via some simulation studies and a real application.",
keywords = "EM algorithm, Model selection, Penalized likelihood, Proportionality constraints",
author = "Hai Liu and Chan, {Kung Sik}",
year = "2010",
month = "7",
language = "English",
volume = "35",
pages = "1--26",
journal = "Journal of Statistical Software",
issn = "1548-7660",
publisher = "University of California at Los Angeles",
number = "11",

}

TY - JOUR

T1 - Introducing COZIGAM

T2 - An R package for unconstrained and constrained zero-inflated generalized additive model analysis

AU - Liu, Hai

AU - Chan, Kung Sik

PY - 2010/7

Y1 - 2010/7

N2 - Zero-inflation problem is very common in ecological studies as well as other areas. Nonparametric regression with zero-inflated data may be studied via the zero-inflated generalized additive model (ZIGAM), which assumes that the zero-inflated responses come from a probabilistic mixture of zero and a regular component whose distribution belongs to the 1-parameter exponential family. With the further assumption that the probability of non-zero-inflation is some monotonic function of the mean of the regular component, we propose the constrained zero-inflated generalized additive model (COZIGAM) for analyzing zero-inflated data. When the hypothesized constraint obtains, the new approach provides a unified framework for modeling zero-inflated data, which is more parsimonious and efficient than the unconstrained ZIGAM. We have developed an R package COZIGAM which contains functions that implement an iterative algorithm for fitting ZIGAMs and COZIGAMs to zero-inflated data based on the penalized likelihood approach. Other functions included in the package are useful for model prediction and model selection. We demonstrate the use of the COZIGAM package via some simulation studies and a real application.

AB - Zero-inflation problem is very common in ecological studies as well as other areas. Nonparametric regression with zero-inflated data may be studied via the zero-inflated generalized additive model (ZIGAM), which assumes that the zero-inflated responses come from a probabilistic mixture of zero and a regular component whose distribution belongs to the 1-parameter exponential family. With the further assumption that the probability of non-zero-inflation is some monotonic function of the mean of the regular component, we propose the constrained zero-inflated generalized additive model (COZIGAM) for analyzing zero-inflated data. When the hypothesized constraint obtains, the new approach provides a unified framework for modeling zero-inflated data, which is more parsimonious and efficient than the unconstrained ZIGAM. We have developed an R package COZIGAM which contains functions that implement an iterative algorithm for fitting ZIGAMs and COZIGAMs to zero-inflated data based on the penalized likelihood approach. Other functions included in the package are useful for model prediction and model selection. We demonstrate the use of the COZIGAM package via some simulation studies and a real application.

KW - EM algorithm

KW - Model selection

KW - Penalized likelihood

KW - Proportionality constraints

UR - http://www.scopus.com/inward/record.url?scp=77955153103&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77955153103&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:77955153103

VL - 35

SP - 1

EP - 26

JO - Journal of Statistical Software

JF - Journal of Statistical Software

SN - 1548-7660

IS - 11

ER -