Session-Based Recommendation

A recommendation approach based on the behavior flow within the current session rather than long-term user history.

Session-based recommendation systems offer major advantages when user identity is unknown or historical interaction is limited. Recommendations are generated from clicks, navigation order, and short-term intent signals within the current session. They are especially critical in platforms with large volumes of anonymous traffic. This approach is strong for short-term intent prediction, though it makes limited use of long-term preference information.