A deep selective learning network for cross-domain recommendation

Huiting Liu, Qian Liu, Peipei Li, Peng Zhao, Xindong Wu

Abstract: Highlights•Profile the preference of users and the characteristic of items from reviews.•Fit for the situation where there are no overlapping entities between two domains.•Propose an automatic selection mechanism to select informative entities.•Use a regularized constraint to avoid a trivial solution and select more entities.•Exploit supervised information to select entities with high predictive confidence.