Session-aware Linear Item-Item Models for Session-based Recommendation

Session-based recommendation aims at predicting the next item given a sequence of previous items consumed in the session. (e.g., e-commerce or multimedia streaming services) Specifically, session data exhibits its unique characteristics, i.e., session consistency, sequential dependency, repeated item consumption, and timeliness of sessions. In this paper, we propose simple-yet-effective session-aware linear models, considering the holistic aspects of the sessions. This holistic nature of our model helps improve the quality of recommendations, and more importantly provides a generalized framework for various session data. Thanks to the closed-form solution for the linear models, the proposed models are highly scalable. Experimental results demonstrate that our simple linear models show comparable or state-of-the-art performance in various metrics on multiple real-world datasets.

• Machine Intelligence

• Data Mining and Modeling