Emotion Classification in Voice Data Using Different Machine Learning Methods

Abstract

Emotional intelligence is essential to many fields, such as psychology, medicine, and human-computer interaction. Using datasets from RAVDESS, SAVEE, CREMA, and TESS, which include a range of emotions such neutrality, surprise, happiness, sorrow, contempt, fury, and fear, the research thoroughly examines machine learning algorithms for emotion recognition in audio data. Long Short-Term Memory (LSTM) networks, decision trees, and convolutional neural networks are the three models that are being studied. CNNs are excellent at extracting characteristics from audio sources, LSTMs find temporal correlations from data, and decision trees provide simple categorization. Performance evaluation uses metrics such as recall, accuracy, precision, and the F1 score. Importantly, the CNN model surpasses Decision Trees and LSTM networks in emotion categorization accuracy, by 73% and 76%, respectively, reaching a staggering 92%. This study provides valuable information regarding how different machine learning models perform when it comes to audio-based emotion recognition. The results of this research will have a significant impact on the development of reliable emotion recognition systems for emotional computing, human-robot interaction, and mental health assessment. Future research could look into ensemble techniques or hybrid models to enhance emotion detection skills and further the development of increasingly intricate and accurate emotion recognition systems.