Spectral-Structure Synergy Optimization (SSSO) Method

A physics-informed spectral variable selection method integrating characteristic spectral lineshapes with structured synergy effects.

This toolbox implements the Spectral-Structure Synergy Optimization (SSSO) method and its refined version SSSO2.

Key Steps:

• Dictionary Learning: Models characteristic spectral lineshapes (CSLs) using a sparse Gaussian dictionary.
• Adaptive Segmentation: Decomposes the spectrum into potential peak structures via the results of variational Bayesian inference.
• Synergy Optimization: Captures synergistic effects among structures using a structure-based bootstrap sampling strategy.
• Intra-structure Pruning: Removes redundant variables within selected structures to refine the feature subset.

Reference:

P. Wu, T. Chen, X. Zou, and H. Li, "A physics-informed spectral-structure synergy optimization (SSSO) method for consistent and interpretable spectral variable selection."