Advancements in Communication and Systems
Speech Emotion Recognition using Gaussian Mixture Model (GMM) and K-Nearest Neighbors (KNN)
Authors: Kirtika Iyer, Abhay Shukla, Kunal Sharma and Maya Varghese
Publishing Date: 26-05-2024
ISBN: 978-81-955020-7-3
Abstract
This paper aimed to propose a novel methodology to improve the accuracy and efficiency of speech emotion recognition, through to the multilingual setting. The paper’s topic was the low precision obtained by the systems for speech emotion recognition, especially in multilingual settings. The research problem was the performance of the existing systems which could achieve merely 72% accuracy in recognizing the correct emotion from speech. The research’s importance was the enhancement of the performance of these systems in order to improve the user experience and the range of its applications in multilingual settings. The paper uses a research methodology with feature extraction methods and machine learning algorithms, such as Mel-frequency cepstral coefficients, zero-crossing rate, harmonic-to-noise ratio, such as Gaussian Mixture Models, and K-Nearest Neighbors. The proposed methodology analysis leads to a major increase in accuracy, attaining the performance of 82% in the complex multilingual environment. Besides, this research paper describes the areas for future research to allow additional improvement and overcome the possible weaknesses of the designed methodology, contributing to the development of the field.
Keywords
Speech Emotion Recognition, Artificial Intelligence, Emotion Identification, Feature Extraction, Machine learning algorithms, Gaussian Mixture Models, k-Nearest Neighbors.
Cite as
Kirtika Iyer, Abhay Shukla, Kunal Sharma and Maya Varghese, "Speech Emotion Recognition using Gaussian Mixture Model (GMM) and K-Nearest Neighbors (KNN)", In: Ashish Kumar Tripathi and Vivek Shrivastava (eds), Advancements in Communication and Systems, SCRS, India, 2024, pp. 443-455. https://doi.org/10.56155/978-81-955020-7-3-39