ERSS-MLT: Emotion Recognition System from Speech Using Traditional Machine Learning Techniques

Speech Emotion Recognition system aims to measure and recognize the emotional state of human interaction (conversations). Vocal expressiveness is crucial in conveying, the partial of emotional information are recognized as nine basic emotions. Emotion recognition system from speech (ERSS) technology is effective in certain application like recognizing emotion from face challenge universally across the languages due to limited datasets. To address this issue, the system introduces traditional machine learning techniques in feature selection algorithm for English emotional speech created through non-actor elicited responses. The dataset, containing audio samples from individuals, are the largest of their kinds in English representation. Its validated using crowd sourcing to ensure the quality to improve ERSS applications in various fields such as healthcare and education. Therefore, this system also describes the various concrete directions that could be explored in order to improve the security of ERSS.