Computational modeling of behavioral tasks: An illustration on a classic reinforcement learning paradigm