Outlining Ensemble K-Nearest Neighbors Approach for Low-Voltage Power Demand Forecasting

Following the ongoing transformation of the European power system, in the future, it will be necessary to locally balance the increasing share of decentralized renewable energy supply. Therefore, a reliable short-term load forecast down to the level of single low-voltage end-consumers might be required. In this work, we outline how to improve the K-Nearest Neighbors forecaster that we have proposed earlier for the low voltage load forecasting. We motivate the usage of the, so called, Ensemble K-Nearest Neighbors indicating a significant improvement by combining the forecasts obtained using different numbers of nearest neighbors.