MIDI music model

Computational music analysis has received extensive research attention in the recent decade. This study explores music as a statistical time series of notes to predict note events in the future given a music excerpt. Instead of using music theory to model the aesthetics of music, this study applies the Holt-Winter Model. The experiment result proves that musical components, such as notes and beats, are statistically related and can be fit by the Holt-Winter Model. The model also shows the ability to predict a specified number of future notes. Such research can be applied to assisting musicians encountering writer´s block.