Short-Term Power Load Forecasting Based on CNN-BiLSTM Optimized by Snow Ablation Algorithm

To enhance the accuracy of short-term power load forecasting, this paper proposes a hybrid prediction model based on the Snow Ablation Optimization (SAO) algorithm, integrating Convolutional Neural Networks (CNN) and Bidirectional Long Short-Term Memory (BiLSTM) networks. First, linear interpolation is used to impute missing values in the raw data, and the data is normalized. Pearson correlation coefficients are then employed to filter influencing factors. Next, CNN is utilized to extract temporal features from historical load sequences, which are input into the BiLSTM to mine internal variations and train the hybrid model. Finally, SAO is introduced to optimize key parameters of the model, alleviating the inefficiency and suboptimality of manual tuning. Experiments using actual California load data show that the model's MAE and RMSE are reduced by 1.26% and 2.80%, respectively, with an R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> improvement of 0.19%, validating the effectiveness of the proposed model.