Accurate load forecasting greatly influences the planning processes undertaken in operation centers of energy providers that relate to the actual electricity generation，distribution，system maintenance as well as electricity pricing This paper exploits the applicability and compares the performance of the FF-DNN and R-DNN models on the basis of accuracy and computational performance in the medium term forecast of electricity load Firstly，a timefrequency feature selection procedure is proposed because the introduced scheme may adequately learn hidden patterns Then，the FF-DNN and R-DNN models are used for the medium term load forecasting Finally，the herein proposed method is used to predict the load in different seasons in the coming year over real datasets gathered in a period of 5 years Overall，our generated outcomes reveal that the synergistic use of TF feature analysis with DNN enables to obtain higher accuracy