Refining analytic approximation based estimation of mixed multinomial probit models by parameter selection

Abstract Applying analytic approximations for computing multivariate normal cumulative distribution functions has led to a substantial improvement in the estimability of mixed multinomial probit models, both terms accuracy and especially computation time. This paper makes contribution by presenting possible way improve estimating model covariances based on idea parameter selection using cross-validation. Comparisons MACML approach indicate that proposed is able recover covariance parameters more accurately, even when there moderate degree independence between random coefficients. The also estimates efficiently, with standard errors tending be smaller than those approach, which can observed means real data case.