Home
Back
Akaike Information Criterion Formula:
| From: | To: |
The Akaike Information Criterion (AIC) is a statistical measure used to compare different mathematical models for a given set of data. It estimates the quality of each model relative to the others, balancing model fit with complexity to prevent overfitting.
The calculator uses the AIC formula:
Where:
Explanation: The AIC rewards goodness of fit (as assessed by the likelihood function) but penalizes models with more parameters to discourage overfitting.
Details: AIC is widely used in model selection across various scientific disciplines. It helps researchers choose the model that best explains the data with the fewest parameters, promoting parsimony in statistical modeling.
Tips: Enter the number of parameters in your model and the maximum likelihood value. Both values must be positive numbers (parameters ≥ 1, likelihood > 0).
Q1: What does a lower AIC value indicate?
A: A lower AIC value suggests a better model, as it indicates a better fit with fewer parameters relative to other models being compared.
Q2: How big of an AIC difference is significant?
A: Generally, a difference of 2-7 points suggests moderate evidence, while differences greater than 10 points indicate strong evidence for the model with the lower AIC.
Q3: Can AIC be used to compare non-nested models?
A: Yes, one of the advantages of AIC is that it can be used to compare both nested and non-nested models, unlike some other model selection criteria.
Q4: What are the limitations of AIC?
A: AIC assumes that the models being compared are fitted to the same data set and that the sample size is sufficiently large. It may perform poorly with very small sample sizes.
Q5: Is there a corrected version of AIC for small samples?
A: Yes, AICc (corrected AIC) is recommended for small sample sizes, where the ratio of sample size to number of parameters is less than 40.