Single-index model selections

We derive a new model selection criterion for single-index models, AIC(C), by minimising the expected Kullback-Leibler distance between the true and c andidate models. The proposed criterion selects not only relevant variables but also the smoothing parameter for an unknown link function. Thus, it is a general selection criterion that provides a unified approach to model se lection across both parametric and nonparametric functions. Monte Carlo stu dies demonstrate that AICC performs satisfactorily in most situations. We i llustrate the practical Use Of AICC with an empirical example for modelling the hedonic price function for cars. In addition, we extend the applicabil ity Of AICC to partially linear and additive single-index models.