This article mainly discusses how to extract the interested information from massive amounts of micro-blogs and recommend right information to user, which is a hot research area in recommendation systems and social networks, too. To solve this problem, a model called Multi-tags Latent Dirichlet Allocation is proposed. Using this model, topics paid attention by users can be mined effectively and the defect of low degree of differentiation for the short blog content is settled. Experiments showed that the tags of user's micro-blog can be figured out with this model which makes users manage their resources at their convenience and others find their needed resources through tags. The results, experimented on real micro-blog data set, indicate that this model works better than traditional model on extracting tags. Standard measuring index Perplexity is applied to this model to estimate the likelihood of new text. If the number of topics is selected appropriately, the accuracy will be raised to almost 10%.
Abstract-The traditional collaborative filtering algorithm cannot response user interest with time and is lack of time effectiveness. These problems lead to poor recommendation quality. On the basis of the neighbor-based collaborative filtering, a fused method of improved similarity and user interest is proposed. To begin with, we compute similarity from global perspectives obtained with Jaccard similarity, local perspectives obtained with Bhattacharyya Coefficient. Furthermore, we adopt the forgetting curve to represent the user interest preference, adding the interest weight to the new similarity method to update user interest. Finally, we make recommendation prediction by calculating similarity using the method. Experimental results on the Movielens datasets demonstrate that our approach has advantages over state-of-the-art methods in terms of both the discovery of user interest preference and providing highly accuracy recommendations.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
334 Leonard St
Brooklyn, NY 11211
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.