A Decision Support System Using Two-Level Classifier for Smart Grid

Today, big data is not only the data scenario with large volume, but also high-speed and changing all the time. Such data streams commonly exist in Smart Grid facilities. Decision tree as one of the most widely-used analysis methods, has been applied in the decision support system for smart grid. This paper proposes a two-level classifier combining cache-based classifier and incremental decision tree learning, instead of the tree inductions using Hoeffding bound. The simulation result shows that the proposed approach has better accuracy. The combined method can handle high-speed data streams collected from power grid units.