Improving Speech Emotion Recognition and Classification Accuracy Using Hybrid CNN-LSTM-KNN Model
Published in International Journal of Research Publication and Reviews, 2024
This study addresses Speech Emotion Recognition (SER), a vital aspect of human-computer interaction, using a hybrid model combining CNN, LSTM, and KNN algorithms. The CNN extracts acoustic features from speech spectrograms, while LSTM captures temporal dependencies. KNN classifies emotional states based on these features. The model, tested on the TESS dataset, achieves high accuracy and robustness, outperforming traditional methods. It holds promise for real-world applications in sentiment analysis, virtual assistants, and more, advancing SER techniques for improved human-computer interaction.
Recommended citation: Hossain, Md Imran, Md Mojahidul Islam, Tania Nahrin, Md Rashed, and Md Atiqur. "Improving Speech Emotion Recognition and Classification Accuracy Using Hybrid CNN-LSTM-KNN Model." Journal homepage: www. ijrpr. com ISSN 2582: 7421.
Download Paper