| Summary: | 碩士 === 國立屏東科技大學 === 資訊管理系所 === 98 === Different recommendation systems have different recommendation mechanisms. It is recommended interesting items to users which through information filtering or clustering. With the rapid development of the internet and the digital technology, cause online music radio to development. Music recommendation systems also provide online music to users. The common music recommendation systems extract features from music, than computing similarity of those features and it will cluster together who have a high degree similarity of features. This approach in a variety of recommendation mechanism has a good effect, but they are ignoring degree of influence between features and users. Different users face to the same features changes, the extent of the impact is not the same. If use the same way to judge, it will have error result. For this reason, we will provide questionnaire to users, then computing user's weight matrix. This result can help us to understand degree of influence between features and users, improving the rate of recommendation accuracy.
|
|---|
