BiGRU-CNN neural network applied to short-term electric load forecasting

Paper aims This study analyzed the feasibility of BiGRU-CNN artificial neural network as a forecasting tool for short-term electric load. model can serve support related to decision-making by companies in energy sector. Originality Despite large amount scientific research this area, literature still searches more assertive models regarding Thus, model, based on layers BiGRU and CNN architecture networks was tested. already proposed used other similar tasks, however, it has not been load forecasting. Research method The code programmed Python using keras package. forecasts all were carried out 10 times until an acceptable statistical sample reached so that future values are close possible reality. Main findings best when compared classical some hybrid networks. Implications theory practice methodology be applied problems. There is evidence combination different provide efficient results than with only one architecture.