Abstract
This study develops an innovative method for analyzing and clustering tonal trends in Chinese Yue Opera to identify different vocal styles accurately. Linear interpolation is applied to process the time series data of vocal melodies, addressing inconsistent feature dimensions. The second-order difference method extracts tonal trend features. We introduce a fuzzy C-means clustering method enhanced by quantum particle swarm optimization (QPSO) to manage data uncertainties, improving classification accuracy and convergence speed. Additionally, we employ a cross-correlation function to eliminate uncertainties from tonal transition redundancies. We designed a detection algorithm using trend data to validate our clustering method, thereby enhancing the accuracy of the analysis of tonal ranges and potential models. This method detects whether Yue Opera adheres to traditional rhythmic norms and models the regularity of musical tones and vocal patterns. Simulation results reveal that our approach achieves a 91.4% accuracy in classifying vocal styles, surpassing traditional methods and demonstrating its potential for identifying various styles. This research offers technical support for Yue Opera music education and interdisciplinary research. The findings enhance the quality of artistic creation and performance in Yue Opera, ensuring its preservation and development.