Estimation of discrete choice models with hybrid stochastic adaptive batch size algorithms

The emergence of Big Data has enabled new research perspectives in the discrete choice community. While techniques to estimate Machine Learning models on a massive amount data are well established, these have not yet been fully explored for estimation statistical Discrete Choice Models based random utility framework. In this article, we provide ways dealing with large datasets context Models. We achieve by proposing efficient stochastic optimization algorithms and extensively testing them alongside existing approaches. develop three main contributions: use Hessian, modification batch size, change algorithm depending size. A comprehensive experimental comparison fifteen is conducted across ten benchmark Model cases. results indicate that HAMABS algorithm, hybrid adaptive size method, best performing benchmarks. This speeds up time factor 23 largest model compared used practice. integration software will significantly reduce required therefore enable researchers practitioners explore approaches specification models.