Short-term Electricity Price Forecasting Based on Data Mining

Abstract

In the electricity market, the relationship between the electricity price and each participant is very close. For the power generation side, the power buyer, and the market regulator, electricity price forecasting has always been an important task. A short-term electricity price forecasting method based on data mining is designed. Based on the data mining technology and load influencing factors, a fuzzy classification miner is designed to implement data mining for short-term electricity price forecasting in the electricity market. It performs transformation processing and normalization process on that mining data. Based on LSTM, a short-term electricity price forecasting model is constructed, which is composed of a hidden layer, input layer, and input gate to implement short-term electricity price forecasting. The case test results indicate that the relative error of the short-term electricity price forecasting results in spring is the lowest, the absolute error of the short-term electricity price forecasting results in autumn is the lowest, and the overall error is relatively low. The minimum time of the electricity price forecasting is about 4756s, and it only takes a relatively short time to complete the short-term price forecasting.