Enhancing Speech Emotional Recognition through a Multi-Layer Perceptron Model DOI

L. Arun Raj,

Inderpaal Singh Rathna Balaji,

Sheik Taaha Hussain

и другие.

Опубликована: Авг. 23, 2023

Detecting emotions in audio speech files is a significant challenge for machines, despite the natural skill of humans this area. While computers excel at comprehending informational content, deciphering underlying remains formidable challenge. At juncture, Speech Emotion Recognition System (SERS) emerges as an important component solution. To reliably categorize data into distinct emotional categories such happiness, sorrow, anger, and neutrality, specific model called SERS was developed. It has wide range potential uses, from healthcare to contact centers. In study, following Machine Learning models were utilized Support Vector (SVM), k-nearest neighbors (KNN), Decision Tree (DT). These yielded lower accuracy rates when categorizing emotions. Thus, improve rate three crucial features Mel, chroma, MFCC are applied by constructed analyzed utilizing grid search-based Multi-Layer Perceptron (MLP) classifier accurate results. The proposed produced 75.97% rate, which quite remarkable.

Язык: Английский

Hybrid Lyrebird Red Panda Optimization_Shepard Convolutional Neural Network for Recognition of Speech Emotion in Audio Signals DOI

N Kanimozhi,

R. Devi Priya

Neurocomputing, Год журнала: 2025, Номер unknown, С. 129506 - 129506

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

0
Enhancing Speech Emotional Recognition through a Multi-Layer Perceptron Model DOI

L. Arun Raj,

Inderpaal Singh Rathna Balaji,

Sheik Taaha Hussain

и другие.

Опубликована: Авг. 23, 2023

Detecting emotions in audio speech files is a significant challenge for machines, despite the natural skill of humans this area. While computers excel at comprehending informational content, deciphering underlying remains formidable challenge. At juncture, Speech Emotion Recognition System (SERS) emerges as an important component solution. To reliably categorize data into distinct emotional categories such happiness, sorrow, anger, and neutrality, specific model called SERS was developed. It has wide range potential uses, from healthcare to contact centers. In study, following Machine Learning models were utilized Support Vector (SVM), k-nearest neighbors (KNN), Decision Tree (DT). These yielded lower accuracy rates when categorizing emotions. Thus, improve rate three crucial features Mel, chroma, MFCC are applied by constructed analyzed utilizing grid search-based Multi-Layer Perceptron (MLP) classifier accurate results. The proposed produced 75.97% rate, which quite remarkable.

Язык: Английский

Процитировано

0