Prediction and Visualization of Electricity Consumption in the Philippines Using Artificial Neural Networks, Particle Swarm Optimization, and Autoregressive Integrated Moving Average

Electricity is vital in the development of a country like the Philippines. The researchers of the study developed a long-term prediction model to predict and visualize the regional electricity consumption in the Philippines through implementing Particle Swarm Optimization (PSO) instead of Backpropagation (BP) to train the Artificial Neural Networks (ANN) and implementing Autoregressive Integrated Moving Average (ARIMA) to forecast future predictors. The average prediction accuracy of BP-ANN is 85.95% while the average prediction accuracy of PSO-ANN is 92.40%. Moreover, the average accuracy of PSO-ANN and ARIMA is 96.05%. The results indicated that PSO-ANN is a better prediction model than BP-ANN and that ARIMA performed well in forecasting the future predictors. The study can be of great help to improve the electricity allocation in the Philippines.

[1]  R. K. Agrawal,et al.  PSO based Neural Networks vs. traditional statistical models for seasonal time series forecasting , 2013, 2013 3rd IEEE International Advance Computing Conference (IACC).

[2]  Ting Wang,et al.  Temperature prediction of electrical equipment based on autoregressive integrated moving average model , 2017, 2017 32nd Youth Academic Annual Conference of Chinese Association of Automation (YAC).

[3]  Hiroyuki Mori,et al.  A hybrid intelligent system for electricity price forecasting , 2016, 2016 IEEE Region 10 Conference (TENCON).

[4]  S. Alipour,et al.  Analysis and predicting electricity energy consumption using data mining techniques — A case study I.R. Iran — Mazandaran province , 2015, 2015 2nd International Conference on Pattern Recognition and Image Analysis (IPRIA).