SGC-Pro: A Symbolic Genetic-like coding Framework for Interpretable Classification of Astronomical Spectra

Astronomical spectroscopy is a cornerstone of modern astrophysics, revealing the chemical makeup, dynamics, and evolution of stars, galaxies, and nebulae. Yet traditional approaches to spectral analysis often rely on large labeled datasets, computationally heavy pipelines, or black-box machine learning models that limit interpretability and transparency. This creates a gap between data-driven accuracy and human-readable understanding. To bridge this gap, we introduce SGC-Pro, a symbolic framework inspired by genetic coding that transforms emission and absorption peaks into short symbolic sequences. By encoding inter-peak distances into discrete symbols, SGC-Pro captures the structural “fingerprints” of spectra in a compact, interpretable, and scalable form. This representation enables spectra to be classified, clustered, and compared without supervised training or complex preprocessing. In tests on 10000 real SDSS spectra—including galaxies, nebulae, and stellar types—SGC-Pro demonstrated competitive accuracy while preserving clarity and scientific interpretability. Beyond classification, the symbolic sequences naturally expose rare or unusual spectral patterns, making SGC-Pro highly promising for anomaly detection, unsupervised exploration, and the discovery of unexpected astrophysical phenomena. Its lightweight nature also makes it suitable for large-scale survey pipelines and educational use. Looking forward, SGC-Pro can be extended with redshift-aware encoding and multi-wavelength support, evolving into a versatile tool for interpretable spectral science. By uniting simplicity with discovery potential, it opens a fresh path toward understanding the universe through symbolic representation.