Design of a Music Recommendation System Based on Look-alike and K-means Algorithms

With the development of internet technology and the arrival of the information age, there have been many algorithms and systems that can push related things based on user preferences. Music playback platforms would adopt a personalized recommendation method to push music that meets users’ tastes to improve their experience and stickiness, so the issue of whether personalized recommendation is accurate becomes the key. The Look-alike algorithm can identify people with potential associations, while the K-means algorithm can classify them based on the potential connections of data, making them fully feasible for application in music recommendation systems. This article designed a music recommendation system by studying the Look-alike and K-means algorithms, and conducted experiments on the designed system. Through experiments, it was proved that the system designed in this article could effectively improve the accuracy of personalized recommendations. In the experiment, the proportion of music types pushed by the system differed by 0.18% from that in the user’s playlist.