An adaptable four-bar mechanism is proposed in this paper to validate data-based modeling methods. The adaptability of the four-bar is reached by using a linear actuator into the crank. This actuator is a small mechanism based on a DC motor and an endless screw. This proposal is used to validate a specific data-based method, based on a non-temporal prediction, which considers that the model parameters have no physical meaning. This method has a better performance than the standard least square filter (RLS), as we experimentally show.