Speech emotion recognition system using MLP Classifier Algorithm

Research on speech emotion recognition (SER) is ongoing and has numerous applications in fields like healthcare, education, and human-computer interaction. The RAVDESS dataset for SER, which includes a varied collection of emotional speech recordings from 24 actors, is examined in this research as a potential resource for SER. Using the RAVDESS dataset, we specifically investigate the effects of several feature extraction methods, machine learning algorithms, and data augmentation strategies on SER accuracy. Our findings demonstrate that our proposed method was able to perform better with higher accuracy than most models, showcasing the RAVDESS dataset's potential for SER research. We also explore the generalizability of our results and contrast them with those from other datasets. This research can offer practical insights for enhancing SER systems' accuracy and it can have a significant impact on the development of emotion-aware systems.