1) To derive the model selection criterion in order to overcome the weak signal-to-noise ratio and to reduce the probability of over/underfitting in univariate regression model, called adjusted penalty information criterion (APIC).
2) To examine the performance of APIC, the proposed model selection criterion, relative to the well-known criteria, AIC, BIC, and KIC, under various circumstances as follows:
Sample sizes ( ) are difference,
Orders of true model ( ) are difference,
Regression coefficients ( ) are difference,
Variances of error terms ( ) are difference,
Distributions of independent variables (Dist. of ) are difference,
Distributions of error terms (Dist. of ) are difference.