A method is provided in one example embodiment and may include training a statistical model representing radio access point loads or load changes for radio access points, wherein the statistical model is trained using, at least in part, historical measurement data associated with previous user equipment (UE) handovers among the radio access points and wherein the historical measurement data used to train the statistical model is gathered before and after the previous UE handovers; collecting current measurement data associated with a source radio access point and a target radio access point; and calculating a predicted load or load change for the target radio access point for one or more potential UE handovers from the source radio access point to the target radio access point for one or more UE based, at least in part, on application of the current measurement data to the trained statistical model.