RNN-based Deep Learning for One-hour ahead Load Forecasting
Bùi Văn Ga, Van Hoa Nguyen, Tung Lam Pham, Joongheon Kim, Yeong Min Jang
Abstract
This paper is focusing on one hour-ahead forecasting on Power Load, using Recurrent Neural Network based scheme. This study only uses the generated data of Kookmin University's Load, so it required a considerable number of resources for forecasting. Multi-scaled RNN model was proposed for the Load Forecasting, which is suitable for both short term and long term memory.
Topics & Concepts
Recurrent neural networkComputer scienceLong short term memoryTerm (time)Artificial intelligenceArtificial neural networkMachine learningScheme (mathematics)Deep learningProbabilistic forecastingProbabilistic logicPhysicsMathematical analysisMathematicsQuantum mechanicsEnergy Load and Power ForecastingTraffic Prediction and Management TechniquesNeural Networks and Applications