Research and Application of Electricity Price Prediction Model Based on CNN-TPA-GRU
This paper introduces a method of electricity price prediction to solve the fluctuation of electricity market price.In this method,Convolution Neural Networks are used to extract local features of input sequences and reduce dimensionality.At the same time,the Temporal Pattern Attention mechanism is applied to consider the dependencies between different time steps,assign weights to each time step,and optimize input characteristics of the Gate Recurrent Unit.The model is verified on TGE dataset,and compares with other algorithms,and reaches the optimal in each evaluation index,which confirmed its adaptability and feasibility in the power spot market.
国家科技期刊平台
NSTL
万方数据
维普
