IJMSTA - Vol. 2 - Issue 1 - January 2020
Evaluation of Deep Learning and Data Mining Techniques for Audio Data Stream Classification
Authors: Hayder K. Fatlawi, Attila Kiss
Abstract - Classification of audio streams into a meaningful category like music genres has increasing research interest due to the need of the multimedia websites categori-zation and user profiling. Audio signals require many of preprocessing and fea-tures extraction operations. In this paper, two types of data mining classifiers (batch and stream) has been used with 40 extracted features while CNN deep learning classifier utilized to classify images that obtained from waveform and spectrogram representations of music data. The results showed that Random Forest classifier had the best performance in both batch and stream classification with accuracy 71% and 74.6% respectively.
Keywords: Audio Stream, Convolution Neural Network, Random Forest Decision Tree, Bayesian model SVM, Music Genres
Cite this paper as
About this paper
Fatlawi, H.K., Attila Kiss, A.: Evaluation of Deep Learning and Data Mining Techniques for Audio Data Stream Classification. IJMSTA. 2020 Jan 7; 2 (1): 10-18.