A Method for Identifying Abnormal Risk Points in Electricity Market Prices Based on Time Series Decomposition

In view of the abnormal point situation of electricity price in the electricity market, this paper proposes an identification method of abnormal point of electricity price based on the residual component after time series decomposition combined with a variety of time series data anomaly identification methods. First, the grid electricity price data is decomposed into time series. Based on the residual components obtained from the decomposition, the Box plot method, Histogram-Based Method, Z-score method, Robust Z-score method and other methods are further used for anomaly detection and identification of abnormal data points. Afterwards, through comparative analysis of the identification results, it was found that the electricity market price abnormal risk point identification method based on time series decomposition combined with the Robust Z-score method has the best effect, far better than the traditional point-by-point identification method.