A Survey on Cross-domain Recommendation: Taxonomies, Methods, and Future Directions

Abstract

Traditional recommendation systems are faced with two long-standing obstacles, namely, data sparsity and cold-start problems, which promote the emergence and development of Cross-Domain Recommendation (CDR). The core idea of CDR is to leverage information collected from other domains to alleviate the two problems in one domain. Over the last decade, many efforts have been engaged for cross-domain recommendation. Recently, with the development of deep learning and neural networks, a large number of methods have emerged. However, there is a limited number of systematic surveys on CDR, especially regarding the latest proposed methods as well as the recommendation scenarios and recommendation tasks they address. In this survey paper, we first proposed a two-level taxonomy of cross-domain recommendation which classifies different recommendation scenarios and recommendation tasks. We then introduce and summarize existing cross-domain recommendation approaches under different recommendation scenarios in a structured manner. We also organize datasets commonly used. We conclude this survey by providing several potential research directions about this field.

Multi Sequential Embedding-based Cross-Domain Recommendation

“…ratings on products of different domains are generally unevenly distributed in real recommendation services, CDRs leverage relatively enriched data between multiple domains to alleviate the two problems in a target domain [25]. Over the last decade, many efforts have been engaged for the CDR, and many pieces of systematical literature [20,44,49] have categorized them in terms of domain, overlap, and recommendation task. With the recent development of deep learning and neural networks, many CDRs based on them have emerged.…”

“…As more and more users interact with several domains, it increases opportunities for leveraging information collected from other domains to alleviate the two problems (i.e., data sparsity and cold-start issues) in one domain. It leads to Cross-Domain Recommendation (CDR) which has attracted increasing attention in recent years [44]. Cross-domain Recommender System (CDRS) aims to leverage all available data from multiple domains to generate better recommendations on a target domain.…”

“…In the era of Big data, people are easily overwhelmed by a tremendous amount of information [16]. Millions of new products, text, and videos are being released rapidly, which makes it difficult for people to find their interested items [44]. Recommender systems are special programs designed to track users' interests and offer items to the users [18].…”

“…Third, it also needs to consider how to transfer the knowledge across domains, besides modeling user interests. The last one leads to two core issues for CDR, namely, what and how to transfer [44]. In terms of the second point above, conventional CDRSs are single-target approaches that leverage the auxiliary information from a richer domain to improve recommendation performance in a sparser one [49].…”

Recommender systems based on graph embedding techniques: A comprehensive review

“…As for the recent applications of bipartite graph embedding-based recommendation based on the above techniques, the cross-domain recommendation [213][214][215] seems to be a promising direction oriented to the alleviation of cold start and sparsity problems by relating two or more recommender systems. In detail, the cross-domain [213] refers to two types of domains: one is the target domain, in which the recommendation implements and the other is the source domain consisting of other recommender systems providing additional observed interactions corresponding to the users and items in the target domain.…”