A RNN based time series approach for forecasting turkish electricity load

RNN, LSTM and GRU variations have been increasing its popularity on time-series applications. Liberalization of Turkish Electricity Market empowers the necessity of better electricity consumption prediction systems. This paper presents a Recurrent Neural Networks (RNN), Long-Short Term Memory (LSTM), Gated Recurrent Units (GRU) based time series forecasting experiments on Turkish electricity load prediction. Resulting %0.71 MAPE success of our experiments yields better results than existing researches based on ARIMA and artificial neural networks on Turkish electricity load forecasting which have %2.6 and %1.8 success rate respectively.

[1]  Yoshua Bengio,et al.  On the Properties of Neural Machine Translation: Encoder–Decoder Approaches , 2014, SSST@EMNLP.

[2]  Wojciech Zaremba,et al.  An Empirical Exploration of Recurrent Network Architectures , 2015, ICML.

[3]  Wei-Chiang Hong,et al.  Electric load forecasting by the SVR model with differential empirical mode decomposition and auto regression , 2016, Neurocomputing.

[4]  Xiaohua Li,et al.  Electric load forecasting in smart grids using Long-Short-Term-Memory based Recurrent Neural Network , 2017, 2017 51st Annual Conference on Information Sciences and Systems (CISS).

[5]  Robert Jenssen,et al.  An overview and comparative analysis of Recurrent Neural Networks for Short Term Load Forecasting , 2017, ArXiv.

[6]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[7]  Sudhansu Kumar Mishra,et al.  A Review of Short Term Load Forecasting using Artificial Neural Network Models , 2015 .

[8]  Göksel Biricik,et al.  Artificial neural network and SARIMA based models for power load forecasting in Turkish electricity market , 2017, PloS one.

[9]  Ronald J. Williams,et al.  A Learning Algorithm for Continually Running Fully Recurrent Neural Networks , 1989, Neural Computation.

[10]  Ismet Erkmen,et al.  Intelligent short-term load forecasting in Turkey , 2006 .

[11]  Abdollah Kavousi-Fard,et al.  A hybrid method based on wavelet, ANN and ARIMA model for short-term load forecasting , 2014, J. Exp. Theor. Artif. Intell..

[12]  Razvan Pascanu,et al.  On the difficulty of training recurrent neural networks , 2012, ICML.

[13]  Vrizlynn L. L. Thing,et al.  PowerLSTM: Power Demand Forecasting Using Long Short-Term Memory Neural Network , 2017, ADMA.

[14]  Pei-Chann Chang,et al.  A multiple time series-based recurrent neural network for short-term load forecasting , 2017, Soft Computing.

[15]  Göksel Biricik,et al.  Power load forecast system for Turkish electric market , 2015, 2015 23nd Signal Processing and Communications Applications Conference (SIU).

[16]  Yuan Yu,et al.  TensorFlow: A system for large-scale machine learning , 2016, OSDI.