Short‐Term Wind and Solar Power Prediction Based on Feature Selection and Improved Long‐ and Short‐Term Time‐Series Networks

Abstract

In terms of the problems of high feature dimension and large data redundancy in the wind and solar power prediction method, an improved prediction model is proposed by combining feature selection methods with the long- and short-term time-series network (LSTNet). The long short-term memory (LSTM) unit in the LSTNet model is replaced with the bidirectional long short-term memory (BiLSTM), which enables recursive response training for the states of hidden layers at the start and end of the sequence. For feature selection, both feature screening and dimension reduction methods are considered, including random forest (RF), grey relational analysis (GRA), and principal component analysis (PCA). Finally, based on wind and solar power data, the effectiveness of the proposed methods is verified, where the RF-LSTNet performs the best. For wind power prediction, the mean absolute percentage error is reduced by 29.7% and root mean square error is reduced by 24.1% compared with the traditional LSTNet model, and for solar power prediction, the MAPE is reduced by 12.9% and RMSE is reduced by 3.8%.

Construction and simulation study of wind power prediction model based on BP neural network

“…In Equation ( 6): E denotes the wind power prediction result error condition (/cost function); k denotes the prediction behavior occurred; num denotes the total amount of data related to wind power; k d used to represent the expected output value of the wind power prediction results; k y denotes the actual prediction output value [9]. According to Equation (6), in the process of training the network on the prediction results, the cost function that is closer to the 0weight value, i.e., the bias value, should be mined, i.e., the expression of the function with the lowest error value should be found.…”

Contextualizing injury severity from occupational accident reports using an optimized deep learning prediction model

“…Furthermore, this type of pipeline represents a novel exploration in predictive modeling and has been gaining traction in related studies. Such approaches have been widely adopted across various domains including clinical data classification ( Kong & Yu, 2018 ; Wu et al, 2020 ), financial analysis ( Ma, Han & Fu, 2019 ; Pai & Ilango, 2020 ), and solar power prediction ( Wang et al, 2023 ). All these related studies acknowledge the effectiveness of combining RF for feature importance and LSTM networks for sequence modeling.…”