Abstract:To accurately predict wood dye color matching formulations, a hybrid neural network model combining Linear Discriminant Analysis (LDA), Improved Cuckoo Search (ICS) algorithm, Convolutional Neural Network (CNN), and Gated Recurrent Unit (GRU) is proposed. The model processes and classifies the spectral information by LDA for dimensionality reduction; extracts essential features utilizing CNN; inputs these characteristics into GRU for training; and optimizes the hyperparameters in the network using the ICS algorithm. The model"s performance is measured through various evaluation criteria, including the coefficient of determination (R2) and the Color Difference Calculation Formula (CIEDE2000). In comparison with multiple traditional models, the proposed model demonstrates superior performance. Additionally, the model has a relatively low number of parameters, high computational efficiency, and excellent stability and reliability. The results indicate that the proposed model exhibits significant advantages when applied to predicting wood dye color matching formulations based on spectral information.