Graphical Local Genetic Algorithm for High-Dimensional Log-Linear Models

Graphical log-linear models are effective for representing complex structures that emerge from high-dimensional data. It is challenging to fit an appropriate model in the setting and many existing methods rely on a convenient class of models, called decomposable which lend well stepwise approach. However, these restrict pool candidate they can search, difficult scale. be shown non-decomposable approximated by its minimal triangulation, thus extending computational properties any model. In this paper, we propose local genetic algorithm with crossover-hill-climbing operator, adapted graphical models. We show used successfully both low number variables high variables. use posterior probability as measure fitness parallel computing decrease computation time.