Systematic Literature Review on Application of Naive Bayes Algorithm for Large Audio Data Classification

Abstract

The increasing volume of audio data in areas like speech recognition, music genre classification, and environment sound analysis has created a need for more effective and scalable classification algorithms. This systematic literature review focuses on the use of the Naive Bayes algorithm for large-scale audio data classification and evaluates 39 peer-reviewed articles published in the last five years. The review analyses how Naive Bayes has been applied to difficulties such as feature extraction, model training, and real-time classification of audio signals considering its simplicity and computational efficiency. We assess its performance against more sophisticated machine learning techniques and its flexibility with pre-processing, ensemble models, and cross-layer control algorithms. Findings demonstrate that although Naive Bayes does not always outperform deep learning algorithms, this remains a strong option when low latency, explainability, and minimal resources are required. The review also points out gaps in existing research and discusses potential approaches to improve the algorithm's performance in audio-based tasks.