Document Type



This item is available under a Creative Commons License for non-commercial use only


Computer Sciences

Publication Details

Dissertation submitted in partial fulfilment of the requirements of Technological University Dublin for the degree of M.Sc. in Computing (Stream), January 2018.


Music Information Retrieval (MIR) focuses on extracting meaningful information from music content. MIR is a growing field of research with many applications such as music recommendation systems, fingerprinting, query-by-humming or music genre classification. This study aims to classify the styles of Western classical music, as this has not been explored to a great extent by MIR. In particular, this research will evaluate the impact of different music characteristics on identifying the musical period of Baroque, Classical, Romantic and Modern. In order to easily extract features related to music theory, symbolic representation or music scores were used, instead of audio format. A collection of 870 Western classical music piano scores was downloaded from different sources such as KernScore library (humdrum format) or the Musescore community (MusicXML format). Several global features were constructed by parsing the files and accessing the symbolic information, including notes and duration. These features include melodic intervals, chord types, pitch and rhythm histograms and were based on previous studies and music theory research. Using a radial kernel support vector machine algorithm, different classification models were created to analyse the contribution of the main musical properties: rhythm, pitch, harmony and melody. The study findings revealed that the harmony features were significant predictors of the music styles. The research also confirmed that the musical styles evolved gradually and that the changes in the tonal system through the years, appeared to be the most significant change to identify the styles. This is consistent with the findings of other researchers. The overall accuracy of the model using all the available features achieved an accuracy of 84.3%. It was found that of the four periods studied, it was most difficult to classify music from the Modern period.