Latent Factor Model Based on Simple Singular Value Decomposition for Personalized Comment Recommendation

With the fast development of the e-commercial and content management web application in web 2.0 communities, more and more web communities support users in making comment about the objects they have reviewed. Comments do assist users to learn about the items they are reviewing. However, there are always hundreds of comments about an item, and to review them one by one is a time consuming job. Since there are some comments are given casually and some are irrelevant to the user. Motivated by this situation, we propose a latent factor model based on singular value decomposition(SVD) for profiling user and comment in order to achieve what we call "Personalized Comment Recommendation". We also conduct experiment on the new proposed model in a real life data set, and the experimental result shows that our implementation achieves an good performance.