Abstract:
"In the modern world, AI has become the main focal point of rising data analysis
technologies. Because of the way ML has opened new paths in data analytics tasks, people have
utilized ML for almost every sector to enhance their performance. Applying neural networks to
the music classification process is one of those highly explored machine learning domains.
Manually classifying music resources can be a very exhausting and laborious task. But with the
ability to apply ML to automate this process has made ML researchers to dig deep and propose
various design solutions to automate this task. But because of newly emerging technologies in
the ML field and inventing new sub domains of music classification, there is still a lot to explore
in this field. One of those above mentioned sub domains would be Sinhala music classification.
Compared to its main domain, this sub domain is very underexplored due to less interest shown
in this domain and the changing nature of the music. In this case the author has set a goal to
design a classification system that enables automating Sinhala music classification by genres
using deep neural networks and music information retrieval. A system like this will provide the
ability to perform Sinhala music classification and prediction tasks to all Sinhala music
enthusiasts without any prior technical knowledge. Due to very little research having been
carried out in this area and no other existing systems found similar to the proposed system, this
will be a valuable contribution to the Sri lankan music domain."
