Technology in the modern world has over-simplified the access to information. At a click of a button we have volumes of music accessible on the Internet. Paradoxically, the abundance of available options has only made music discovery and recommendations a complex problem to solve. With huge collections of songs in the online digital libraries, finding a song or an artist is not a problem. However, an actual problem is what to look for that will intuitively satisfy a user’s need. There exists multitude of recommendation algorithms, but many of them do not consider the contextual information in which a user listens to a song. This information is not quantifiable, but it needs to be extracted by some methods so as to provide an additional facet to music recommendations. There is active research in music recommendation to identify various factors that can influence the choice of a song. Songs that are often played together have some inherent correlations between them which at first, does not seem obvious. Thus, an approach is proposed that can extract information using a linear algebraic approach and generate context- aware music recommendations.
Mehta, Ruchit V., "Session Based Music Recommendation using Singular Value Decomposition (SVD)" (2012). Master's Projects. 237.