Application of outlier mining in power load forecasting

According to the theory of power load forecasting, data mining based on historical data of power load is used in load predicting. During the practical application, there are some errors in the data collection, and a load forecasting curve often contains big jagged edges. This paper presents a new outlier data mining approach. It finds sharp angle points between two lines, which correspond to outliers of power load. We smooth the curve at same time outliers are handled. Experiments show that after the new outlier mining approach was applied, load forecast results were improved significantly.