A gradient boosting machine-based framework for electricity energy knowledge discovery

Knowledge discovery in databases (KDD) has an important effect on various fields with the development of information science. Electricity energy forecasting (EEF), a primary application KDD, aims to explore inner potential rule electrical data for purpose serve electricity-related organizations or groups. Meanwhile, advent society attracts more and scholars pay attention EEF. The existing methods EEF focus using high-techs improve experimental results but fail construct applicable electricity KDD framework. To complement research gap, our study propose gradient boosting machine-based framework prediction enrich knowledge applications. be specific, we draw traditional process techniques make reliable extensible. Additionally, leverage Gradient Boosting Machine (GBM) efficiency accuracy approach. We also devise three metrics evaluation proposed including R-square (R2), Mean Absolute Error (MAE), Percentage (MAPE). Besides, collect consumption (EEC) as well meteorological from 2013 2016 New York state take EEC State example. Finally, conduct extensive experiments verify superior performance show that model achieves outstanding (around 0.87 R2, 60.15 MAE, 4.79 MAPE). Compared real value official model, approach remarkable ability. Therefore, find is feasible could provide practical references other types KDD.