The Akaike information criterion: Background, derivation, properties, application, interpretation, and refinements
Top Cited Papers
- 14 March 2019
- journal article
- research article
- Published by Wiley in WIREs Computational Statistics
- Vol. 11 (3)
Abstract
The Akaike information criterion (AIC) is one of the most ubiquitous tools in statistical modeling. The first model selection criterion to gain widespread acceptance, AIC was introduced in 1973 by Hirotugu Akaike as an extension to the maximum likelihood principle. Maximum likelihood is conventionally applied to estimate the parameters of a model once the structure and dimension of the model have been formulated. Akaike's seminal idea was to combine into a single procedure the process of estimation with structural and dimensional determination. This article reviews the conceptual and theoretical foundations for AIC, discusses its properties and its predictive interpretation, and provides a synopsis of important practical issues pertinent to its application. Comparisons and delineations are drawn between AIC and its primary competitor, the Bayesian information criterion (BIC). In addition, the article covers refinements of AIC for settings where the asymptotic conditions and model specification assumptions that underlie the justification of AIC may be violated.This article is categorized under: Software for Computational Statistics > Artificial Intelligence and Expert Systems Statistical Models > Model Selection Statistical and Graphical Methods of Data Analysis > Modeling Methods and Algorithms Statistical and Graphical Methods of Data Analysis > Information Theoretic MethodsKeywords
This publication has 25 references indexed in Scilit:
- Model Selection for Multivariate Regression in Small Samples1994
- A CORRECTED AKAIKE INFORMATION CRITERION FOR VECTOR AUTOREGRESSIVE MODEL SELECTIONJournal of Time Series Analysis, 1993
- Regression and time series model selection in small samplesBiometrika, 1989
- Non-Gaussian State-Space Modeling of Nonstationary Time Series: RejoinderJournal of the American Statistical Association, 1987
- Asymptotically Efficient Selection of the Order of the Model for Estimating Parameters of a Linear ProcessThe Annals of Statistics, 1980
- Estimating the Dimension of a ModelThe Annals of Statistics, 1978
- A new look at the statistical model identificationIEEE Transactions on Automatic Control, 1974
- The Relationship Between Variable Selection and Data Agumentation and a Method for PredictionTechnometrics, 1974
- Some Comments on C PTechnometrics, 1973
- On Information and SufficiencyThe Annals of Mathematical Statistics, 1951