Online recommendations based on dynamic adjustment of recommendation lists

The flourishing of the Internet has increasingly promoted the rise of new types of online news websites with e-commerce portals. Online news websites provide specific information on such topics as lifestyle, fashion news, and a variety of other activities. The provision of online recommendation of activities associated with online news websites has the potential to attract more users and create more benefits. Such online recommendations represent an important online trend. Furthermore, dynamically adjusting recommendation lists to increase users’ click-through rates is important for limited online recommendation layouts; however, existing studies have not addressed this online recommendation issue. This research proposes a novel approach for the dynamic adjustment of recommendation lists to tackle the issue of limited recommendation layouts, and then develops novel online recommendation methods. This research designs novel methods based on non-negative matrix factorization (NMF) and latent Dirichlet allocation (LDA) to predict user preferences for activities, using analysis of browsing news and attending activities. We propose a novel online activity recommendation approach, taking into consideration the interest scores and push scores, for dynamically adjusting the recommendation list. The Most Frequently Pushed (MFP) strategy gives priority to replacing the most frequently pushed activity, while the Not Frequently Clicked (NFC) strategy gives priority to replacing the not frequently clicked activity. We implement our proposed approach on an online news website and evaluate its online recommendation performance. The results of our experiment demonstrate that our proposed approach can enhance the effectiveness of recommendations.