BiLSTM Short-term Wind Power Prediction Based on Attention Mechanism

Abstract

Aiming at the problem that long distance in time dimension will lead to information weakening when processing time-series data, this paper proposes a short-term wind power prediction model based on attention mechanism and bidirectional long short-term memory neural network (BiLSTM-Attention). BiLSTM can obtain the information of the first part and the latter part of the sequence at a time point to extract the effective information contained in the time series data. The attention mechanism can capture important information from effective information, so that the model can better retain the information contained in the data. The example analysis results show that compared with other methods, the proposed method has higher accuracy in predicting wind power, and it is expected to provide some reference for the research in the field of wind power prediction.