NORMA eResearch @NCI Library

Impact of Missing Data on Audio Genre Classification using Convolutional Neural Network

Sathe, Raunak Milind (2022) Impact of Missing Data on Audio Genre Classification using Convolutional Neural Network. Masters thesis, Dublin, National College of Ireland.

[thumbnail of Master of Science]
PDF (Master of Science)
Download (1MB) | Preview
[thumbnail of Configuration manual]
PDF (Configuration manual)
Download (1MB) | Preview


Genre classification of audio files has many applications such as recommendation systems, Digital Signal Processing. Missing data in audio files is very common in real world applications. In this project a novel approach has been implemented to study the impact of missing data on genre classification using deep learning. Convolutional Neural Network is a very powerful deep learning algorithm which is very popular in classification tasks. The dataset used is the GTZAN dataset containing 1000 audio samples from 10 genres. In this project, three experiments were run, first the CNN model using the VGG16 architecture was trained on the GTZAN dataset and then two models were implemented on the two processed GTZAN dataset where segments were removed from the dataset to make new, shorter datasets. Surprisingly, the model showed no downgrade in performance on the processed dataset and thus, the conclusion was reached that missing data need not be very crucial in achieving genre classification.

Item Type: Thesis (Masters)
Uncontrolled Keywords: CNN; VGG16; Genres; Recommendation; Missing Data; GTZAN
Subjects: M Music and Books on Music > M Music
Q Science > QA Mathematics > Electronic computers. Computer science
T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science
Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions: School of Computing > Master of Science in Data Analytics
Depositing User: Tamara Malone
Date Deposited: 10 Mar 2023 16:58
Last Modified: 10 Mar 2023 16:58

Actions (login required)

View Item View Item