Which information criterion uses double the number of parameters in its evaluation?

The Akaike Information Criterion (AIC) is designed to assess the quality of statistical models for a given set of data, and it does so by including a penalty for the number of parameters in the model. Specifically, AIC uses a formula that incorporates the likelihood of the model and adds a complexity penalty based on double the number of estimated parameters. This means that as the number of parameters increases, the penalty in the AIC increases, helping to prevent overfitting by discouraging overly complex models.

In contrast, other criteria such as the Schwarz Bayesian Information Criterion (BIC) apply a larger penalty on the number of parameters, which affects model selection differently. The mean absolute error and mean percentage error measures are not information criteria at all; instead, they focus on assessing predictive accuracy without the model selection aspect that AIC and BIC include. Thus, AIC is the correct option that uniquely identifies with the characteristic of using double the number of parameters in its evaluation process.