---

What music do I like?

Issues

The first step to discovering music you will like is by describing the music you already like. In defining a user's taste in music, there are a number of concerns to take into account. An ideal method to describe a user's taste in music would be simple, flexible, and dynamic, yet also highly specific and quantifiable. The method should be easy to use but robust enough to be used to provide optimal recommendations to the user.

Methods

The most fundamental way to describe the music you like is simply to list your favorite music by one or more of the three fundamental units of music: tracks, albums, and artists. A simpler and more general way of describing the music you like is by naming genres of music which you prefer. A more flexible and dynamic way of describing the music you like is to tag your music using descriptions of or relating to the music. However, the statistically optimal method of describing the music you like is defined in a personalized recommendation engine, which revolves around a statistical understanding of your likes and dislikes. Once the problem of defining what music you like is solved, you must also address the problem of teaching a computer what music you like. Playlists, playcounts, and user-inputted ratings are the three primary methods used to analyze a person's music preferences, and each has its pros and cons.