A Neuro-Fuzzy Based Approach for Energy Consumption and Profit Operation Forecasting

In recent years, the massive growth in the scale of data is being a key factor in the needed data processing approaches. The efficiency of the algorithms of knowledge extraction depends significantly on the quality of the raw data, which can be improved by employing preprocessing techniques. In the field of energy consumption, the forecasting of power cost needed plays a vital role in determining the expected profit. To achieve a forecasting with higher accuracy, it is needed to deal with the large amount of data associated with power plants. It is shown in the literature that the use of artificial neural networks for the forecast electric power consumption and show short term profit operation is capable of achieving forecasting decisions with higher accuracy. In this research work, a neuro-fuzzy based approach for energy consumption and profit operation forecasting is proposed. First, the main influential variables in the consumption of electrical energy are determined. Then, the raw data is pre-processed using the proposed fuzzy-based technique. Finally, an artificial neural network is employed for the forecasting phase. A comparative study is conducted to compare between the proposed approach and the traditional neural networks. It is shown that the achieved forecasting accuracy of the proposed technique is better than what achieved by employing only the neural network.