Wild horse optimizer for variable selection in partial least squares spectral quantification of complex samples
作者：Shaohan Wei#, Yajing Yan#, Ruoxin Wang, Haiyan Bao, Xihui Bian*
關(guān)鍵字：Spectral analysis; Variable selection; Wild horse optimizer; Partial least square
論文來(lái)源：期刊
具體來(lái)源：Applied Sciences, 2026, 15
發(fā)表時(shí)間：2026年
Spectral analysis technology has emerged as vital tools for quantifying complex samples because of its simplicity and high efficiency. However, spectral data has high-dimensional characteristic and traditional variable selection methods struggle to balance computational efficiency and prediction accuracy. Hence, discretized wild horse optimizer (WHO) algorithm was introduced in this study. Firstly, transfer functions were introduced to solve the discrete optimization in spectral variable selection. The optimal number of latent variables (LVs) in partial least squares (PLS), DWHO iterations and the population size were determined to establish the DWHO-PLS quantitative analysis model. Then, three spectral datasets of pork, marzipan and DOSY samples were used to assess the effectiveness of the method. Finally, DWHO-PLS model was compared with full-spectrum PLS, uninformative variable elimination (UVE) -PLS, Monte Carlo (MC)-UVE-PLS, randomization test (RT) -PLS, grey wolf optimizer (GWO) -PLS and whale optimization algorithm (WOA) -PLS. Results show that the number of selected variables by DWHO-PLS was smallest and root mean squared errror of prediction (RMSEP) was the fewest vulue compared with other methods for three datasets. The research indicates that DWHO can effectively simplify the PLS model while enhancing its accuracy and stability.