AIC
The Akaike information criterion is a model-selection score that balances goodness of fit against model complexity, with lower values generally preferred among comparable models.
Effective Data Analysis
AIC is an information criterion that evaluates a statistical model by balancing goodness of fit against model complexity.
- How it is used
- Uses AIC to compare model iterations and assess fit–complexity trade-offs when adding predictors or transforming variables.
- How deep it goes
- Defines relative interpretation and compares AIC with BIC, without deriving the criterion.
- Examples
- Uses housing-price regression with an added floor-count predictor and a product-margin modeling exercise.
- What you can do after
- Evaluate and understand candidate models using AIC alongside other diagnostics.
Statistics Every Programmer Needs
AIC is an information criterion that balances a model's log-likelihood against a penalty for the number of estimated parameters.
- How it is used
- Compares competing time-series models by balancing goodness of fit against parameter-count complexity.
- How deep it goes
- Defines the fit-complexity trade-off and comparison rule, but does not derive the criterion or validate its assumptions in detail.
- Examples
- Uses fitted ARIMA and Holt–Winters time-series models, including printed AIC output for Holt–Winters.
- What you can do after
- Evaluate model alternatives by reading AIC values and preferring the lower value in a same-data comparison.
Statistics Slam Dunk
The Akaike information criterion is a model-selection score that balances goodness of fit against model complexity, with lower values generally preferred among comparable models.
- How it is used
- Compares competing regression and ANOVA models when selecting a useful balance between fit and predictor count.
- How deep it goes
- Defines the score and interpretation rule, then applies it to regression and ANOVA comparisons without deriving its formula.
- Examples
- Uses NBA win regressions and offense-versus-defense ANOVAs, including R outputs with AIC values of 514 versus 518 and 729.2831 versus 760.9676.
- What you can do after
- Evaluate competing models with R's glance() or AIC() functions and interpret lower scores alongside diagnostics and other statistics.
Timeless Algorithms, The Foundational Ideas
The Akaike information criterion is a model selection criterion that estimates relative information loss while balancing fit against model complexity.
- How it is used
- Compares competing models when they are not nested and likelihood-ratio tests are inappropriate.
- How deep it goes
- Explains the maximized-likelihood and complexity-penalty structure and interprets lower scores, without supplying a formula.
- What you can do after
- Evaluate competing models by balancing maximized likelihood against a complexity penalty and preferring lower AIC values.
Read how Timeless Algorithms, The Foundational Ideas explains AIC →
Also covered in
What to know first
- likelihoodA likelihood is the probability of observed data under a model, viewed as a function of the model’s parameters or candidate hypotheses.
- overfittingOverfitting is a modeling condition in which a model adapts too closely to particular data, including its noise or idiosyncrasies, and consequently generalizes poorly to new data.
Related concepts
- model selectionmeasures · Model selection is the process of choosing among candidate models or configurations according to evaluation criteria and suitability for a task.
- linear regressionmeasures · Linear regression is a statistical method that models a numeric response as a linear function of one or more predictor variables with fitted parameters.
- goodness-of-fit measuremeasures · A goodness-of-fit measure quantifies how closely a statistical model or representation agrees with observed data or specified relationships.
- metricmeasures · A metric is a quantitative measure that assigns values to a property or outcome so it can be assessed, compared, or monitored.
- BICcontrasts with · The Bayesian information criterion is a model-selection score that balances likelihood-based goodness of fit against model complexity using a penalty that grows with sample size.
Questions about AIC
What is AIC?
The Akaike information criterion is a model-selection score that balances goodness of fit against model complexity, with lower values generally preferred among comparable models.
Which Manning books explain AIC?
Effective Data Analysis, Statistics Every Programmer Needs, Statistics Slam Dunk and Timeless Algorithms, The Foundational Ideas.
What should I know before learning AIC?
Manning books that teach AIC usually introduce likelihood and overfitting first.
What is the difference between AIC and BIC?
The Akaike information criterion is a model-selection score that balances goodness of fit against model complexity, with lower values generally preferred among comparable models. The Bayesian information criterion is a model-selection score that balances likelihood-based goodness of fit against model complexity using a penalty that grows with sample size.
What is AIC a kind of?
AIC is a kind of information criterion. An information criterion is a quantitative model-selection score that balances goodness of fit against model complexity.
What is AIC also called?
AIC is also called Akaike information criterion.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.



