Mixed Information Flow for Cross-Domain Sequential Recommendations

Cross-domain sequential recommendation is the task of predict next item that user most likely to interact with based on past behavior from multiple domains. One key challenges in cross-domain grasp and transfer flow information domains so as promote recommendations all Previous studies have investigated behavioral by exploring connection between items different The knowledge (i.e., domains) has far been neglected. In this article, we propose a mixed network for consider both incorporating unit . proposed able decide when should be used and, if so, which enrich sequence representation according users’ current preferences. Extensive experiments conducted four e-commerce datasets demonstrate improve performance modeling flow. focus application s scenario two domains, but method can easily extended