Speech Emotion Recognition: A Machine Learning Framework Utilizing MFCC Features and the RAVDESS Dataset

Abstract

The aim of this paper is to design and evaluate a machine learning framework for speech emotion recognition using the RAVDESS dataset. This framework uses Mel Frequency Cepstral Coefficients (MFCC) as audio features to identify and categorize emotions such as happiness, sadness, anger, and surprise. It uses comparative analyses on a host of machine learning algorithms: SVM, Logistic Regression, Random Forest, and Decision Tree to discern their ability towards classifying the emotion. This indicates that, based on SVM models, this classification shows its peak performance in having a high accuracy at about 85.4%.