Author(s) :

Haochen Sun.

Volume/Issue :

Abstract :

This study presents a quantitative algorithmic analysis to classify musical emotion by comparing the predictive power of tempo and key. The work focuses on computational labeling rather than human emotional perception or listener experience. Results show that both tempo and tonality contribute to algorithm-based emotion classification, with tonality (key) yielding higher predictive accuracy than tempo. Statistical testing confirms this difference is highly significant (p < 0.001). Major keys are assigned positive algorithmic emotion scores, while minor keys receive negative scores, independent of tempo. These findings offer a theoretical foundation for music psychology and potential implications for music‑based interventions, including music therapy for children with autism. These results suggest exploratory directions for future therapeutic design but do not constitute evidence of clinical efficacy. This study provides an algorithmic baseline for future work and highlights the relative importance of key and tempo in computational music emotion classification.