Wind Power Prediction Model Based on WOA-BiLSTM-Attention.
2023 IEEE 6th International Conference on Information Systems and Computer Aided Education (ICISCAE)(2023)
摘要
A two-way long and short-term memory network with attention mechanism (WOA-BiLSTM-Attention) model under the whale optimization algorithm is chosen for wind power prediction evaluation using historical data from a wind farm in the context of the high stochastic, fluctuating, and intermittent power generation pattern of wind power. The BiLSTM-Attention model can minimize the loss of previous data and increase effects of significant information. Based on this, the whale optimization algorithm (WOA) is then used for hyperparameter selection to reduce human interference. Finally, comparing the WOA-BiLSTM-Attention model’s with BP, LSTM, BiLSTM, and BiLSTM-Attention, the RMSE using the WOA-BiLSTM-Attention method, compared with BP, LSTM, BiLSTM, and BiLSTM-Attention, is reduced by 1151.9931w, 540.5077w, 353.6318w, 157.9766w, and R
2
increased by 0.0267, 0.012, 0.0074, and 0.003, respectively.This indicates that the WOA-BiLSTM-Attention model is more accurate in wind power prediction results, which provides some prediction for other fields borrowing significance.
更多查看译文
关键词
BiLSTM,WOA,attentional mechanism,wind power prediction
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要