Distinctive Approach for Speech Emotion Recognition Using Machine Learning

Emotions have an important role in human-computer interaction. It improves the efficiency of Human-Computer Interaction (HCI) applications. Speech Emotions Recognition (SER) modules are critical in helping robots in interpreting data through the use of emotions. With precision, vigorousness, and dormancy taken into account, this is a difficult assignment. The proposed methodology in this experiment is to implement two approaches on the same dataset and compare the efficiency. The classification strategy employed is to classify data by implementing four different machine learning algorithms. The Support Vector Method (SVM) model has a 99.57% accuracy on the Tess Dataset and a 63.73% accuracy on the Ravdess Dataset. According to the overall analysis, the model performs better without augmentation, and the model that produced the best possible performance is the Support Vector Machine Algorithm (SVM). Two more potential additions to the model's capabilities are mood swings and depression. Psychologist may implement these techniques to monitor their patients' anxiety attacks. These additions can be made for analyzing emotions of depressed person. It can be advantageous when dealing with a person's mental state.