Online Product Recommendations based on Diversity and Latent Association Analysis on News and Products.

Integrating news websites with product recommendation can create more benefit and is an important trend of online worlds. The information offered by the websites is becom-ing even more complicated. Accordingly, it is important for the websites to implement online recommendation methods that can raise the users' click-through rates and loyalty. In this work, we proposed a novel online product recommendation approach for recom-mending products during news browsing. The proposed method combines online hybrid interest analysis and recommendation diversity. There are cold-start and data sparsity is-sues on the website. Accordingly, a hybrid of collaborative filtering and content-based approach is used to alleviate the issues. Specifically, latent association analysis is con-ducted on user browsing news and products to discover the latent associations between products and news. Moreover, a hybrid method is proposed based on Matrix Factorization and Latent Topic Modeling to predict user preferences for products. In addition, online interest analysis is integrated to adjust users' online product interests according to the cur-rently browsing news. Finally, the proposed approach combines recommendation diversity and users' online interests to raise the chance of discovering potential user preferences on products and enhance the click through rate of online product recommendations. Online evaluations are conducted on a news website to evaluate the proposed approach. Our online experimental results indicate that the proposed approach can enhance the click-through rate of online product recommendations.