Thermal investigation of water-based radiative magnetized micropolar hybrid nanofluid flow subject to impacts of the Cattaneo–Christov flux model on a variable porous stretching sheet with a machine learning approach