A SAX-based method for extracting features of electricity price in power markets

This paper proposes a new method for extracting features of electricity price in power markets The proposed method is based on SAX (symbolic aggregation approximation) of time-series data conversation. Under the deregulated and competitive power markets, it is important to extract the features of electricity price. To understand the complexity of electricity price behavior, this paper proposes a stream mining method with SAX that transforms the data into a symbol of alphabet. It is useful for efficient dimensionality reduction of time series and distance measures. The proposed method classifies time-series data into clusters with symbolic representation, and extracts the features from each cluster. The technique clarifies complicated data in a simply way of symbol representation. The effectiveness of proposed method is demonstrated for real market data of PJM.