Parametric study of hydrogenic inventory in the ITER divertor based on machine learning