Akaike

Akaike is a Japanese surname most notably associated with Hirotugu Akaike, a statistician who developed the Akaike Information Criterion (AIC).

The AIC is a method for model selection that balances goodness of fit with model complexity. It is defined as AIC = -2 log L + 2k, where L is the maximum likelihood of the model and k is the number of estimated parameters. Models with lower AIC values are preferred, reflecting less information loss relative to the true data-generating process. The AIC is grounded in information theory and relates to an estimate of the distance, via the Kullback-Leibler divergence, between the true model and a candidate model.

Since its introduction in 1974, the AIC has become widely used across disciplines, including econometrics, ecology, psychology, biology, and machine learning. It is implemented in many statistical software packages and used to compare competing model structures, not to test hypotheses in isolation.

Variants such as AICc (the small-sample corrected AIC) address bias in small data sets, and multi-model inference uses AIC weights to compare several models concurrently. AIC does not provide an absolute measure of model quality and its values depend on the set of candidate models considered; it may favor more complex models when sample sizes are small. It also contrasts with criteria like BIC, which imposes a heavier penalty for additional parameters.

The name Akaike also denotes a surname in Japan; the information criterion bears Hirotugu Akaike’s name as a tribute to his contributions.