SUPPORT VECTOR MACHINES IN THE LIBERALIZED ENERGY MARKET

摘要

The European Union funds the research project INNOPSE -Innovation Studio and Exemplary Development for Product Service Engineering. Using best practice case studies throughout Europe for creating and managing innovations one example and product service focuses on the simulation of energy market with multi-agent systems. The development of new IT technologies for an efficient and reliable energy market follows the needs by the liberalised electricity trading market in Europe. Data mining with support vector machines (SVMs) is not brand new and it was used in power systems engineering for load forecasting, security and stability classification and some other applications. Here, we present an application of SVMs for modelling real life data for the spot and forward prices (base and peak). The paper shows that SVMs can model trends very reliable, while higher order fluctuation of the true prices is filtered out.