---

Similar Music Networks

Methodology

Based on previously collected information about a listener's preferences, a similar music network looks at the frequency or tendency which which tracks, albums, or artists are related to determine which music listeners deem as similar. This information then forms a similar music network in which similar music collected with a relative amount of strength, and this information can be used to recommend music to listeners. This is a collaborative filtering method that works because, for example, if a listener likes three artists that each are strongly linked to a fourth artist, the listener is very likely to enjoy that artist.

Pros and Cons

A similar music network is an easy-to-understand method of expressing how different music relates to each other. Similar music networks can be formed around each of the three fundamental units of music, and each style of network is valuable for use in a personalized music recommendation engine. However, this system sometimes is too general. If a listener likes one artist well outside of his/her normal range of music interest, the listener most likely appreciates this artist for different reasons or in a different light that listeners whose tastes correlate more strongly to the artist's genre. Therefore, a similar artist network generated by fans of that artist would be far less valuable than a network generated only by fans of that artist who also have a strong correlation of appreciation for the listener's other preferences in music.