Modeling and simulation of bacterial foraging variants: acoustic feature selection and classification

The field of human–computer interaction greatly benefits from the significant role of speech emotion recognition (SER), which finds applications across various domains. However, practical applications of SER still face certain challenges. One such challenge is the variation in emotional expressions among individuals, while another issue arises from the presence of indistinguishable emotions, which can impact the stability of SER systems. This study investigates the application of variants of the Bacterial Foraging Optimization Algorithm (BFOA) in the domain of SER. Experiments are conducted on multiple emotion datasets, including Emo-DB, SAVEE, and SUBESCO, to evaluate the effectiveness of the proposed variants. The findings of this study emphasize the potential of BFOA variants as powerful tools for SER.