Publications by authors named "Zhuo Ying Yang"
A hybrid variable selection and modeling strategy for the determination of target compounds in different spectral datasets.
Spectrochim Acta A Mol Biomol Spectrosc
July 2022
Article Synopsis
- The paper introduces a hybrid technique that combines interval partial least squares (iPLS) and gradient descent (GD) to create quantitative models for analyzing target compounds in different spectral data.
- The iPLS method is utilized for selecting the best variables, while the GD algorithm is employed to build models with these selected variables, optimizing factors like interval number and learning rate for better accuracy.
- The effectiveness of this new iPLS-GD method is tested on three spectral datasets (NIR, H NMR, and EEM), showing it outperforms traditional methods in both modeling and predictive capabilities for identifying compounds in complex samples.