Eigenmood Space: Uncertainty-Aware Spectral Graph Analysis of Psychological Patterns in Classical Persian Poetry
A new AI framework analyzes 10 classical poets with uncertainty-aware multi-label annotation, abstaining on 22.2% of verses.
Researchers Kourosh Shahnazari, Seyed Moein Ayyoubzadeh, and Mohammadali Keshtparvar developed 'Eigenmood Space,' an uncertainty-aware computational framework for psychological analysis of classical Persian poetry. It automatically annotates 61,573 verses across 10 poets with psychological concepts, abstaining on 22.2% where evidence is insufficient. The system uses spectral graph analysis to create poet-level psychological profiles and quantify individuality, enabling scalable digital humanities research while preserving interpretive caution through propagated uncertainty metrics.
Why It Matters
Provides a reproducible, auditable method for large-scale literary analysis while acknowledging AI's limitations in subjective interpretation.