Forecasting and Evaluation Electricity Loss in Thailand via Flower Pollination Extreme Learning Machine Model

Abstract

In many developed countries analyzed and evaluated Loss electricity by using Machine Learning to improve the best solution in forecasting. In Thailand, Provincial Electricity Authority was analyzed in both of technical loss and nontechnical loss to handle the critical situation which could be consequences. In this research, Loss electricity forecasting was studied and analyzed with Machine Learning and Evolutionary Computational for high accuracy when forecasting by using Extreme Learning Machine merged with Flower Pollination Algorithm to find the best solution and use in the real-world. Experiment results show that average of Root Mean Square Error of the proposed model compared with actual data of North, Northeast, Center and South loss electricity datasets from November to December 2017 were 0.5963.