Bootstrap-adjusted quasi-likelihood information criteria for mixed model selection

We propose two model selection criteria relying on the bootstrap approach, denoted by QAICb1 and QAICb2, in framework of linear mixed models. Similar to justification Akaike Information Criterion (AIC), proposed QAICb2 are proved as asymptotically unbiased estimators Kullback–Leibler discrepancy between a candidate true model. However, they defined quasi-likelihood function instead likelihood proven be equivalent. The constructed bias estimation term which method is adopted improve for caused using estimate simulations across variety settings conducted demonstrate that outperform some other existing selecting Generalized estimating equations (GEE) utilized calculate simulations. effectiveness also demonstrated an application Parkinson's Progression Markers Initiative (PPMI) data.