Taxonomy-Oriented Recommendation towards Recommendation with Stage

Abstract

Abstract.In some E-commerce recommender systems, a special class of recommendation involves recommending items to users in a life cycle. For example, customers who have babies will shop on Amazon within a relatively long period, and purchase different products for babies within different growth stages. Traditional recommendation algorithms cannot effectively resolve the situation with a life cycle. In this paper, we model users' behavior with life cycles by employing hand-crafted item taxonomies, of which the background knowledge can be tailored for the computation of personalized recommendation. In particular, our method first formalizes a user's long-term behavior using the item taxonomy, and then identify the exact stage of this user. By incorporating collaborative filtering into our method, we can easily provide a personalized item list to this user through other similar users within the same stage. An empirical evaluation conducted on a purchasing data collection obtained from Amazon demonstrates the efficacy of our proposed method.

TaxoPublish: Towards a solution to automatically personalize taxonomies in e-catalogs

“…similarity, in accordance to the user-user similarity introduced inLi et al (2012). It combines the average mean of Jaccard similarities between all epochs and its Cosine similarity.…”

Next Generation of Recommender Systems: Algorithms and Applications

“…This chapter studies the problem of interest dynamics in a recommender system, and introduces two previous work [LZL11,LHL12] along this direction.…”

“…In this chapter, we start by introducing the problem of interest dynamics in a recommender system, and then discuss two previous work [LZL11,LHL12] published during my Ph.D. study along this direction. In the following, I will present the details of these two approaches, and discuss how the algorithms can incorporate temporal information when modeling user profiles.…”

“…In my previous work, I study the temporal dynamics of user interests in news recommendation [LZL11] and product recommendation [LHL12]. Specifically, in news recommendation, a user's interest evolves over time, depending on casual events happening everyday.…”