Lygame25

Lygame dataset contains two subsets (lygame1and ygame2). The dataset is specifically designed for click-through rate (CTR) prediction under sparse and structure-constrained conditions. Compared with existing datasets like Movielens and Criteo, this is a rather small dataset due to the nature of available resources in the simulation scenario. Unlike many-to-many user-item interactions, Lygame exhibits a strict one-to-many structure. This results in an asymmetric and highly sparse interaction matrix, posing unique challenges for recommendation models. (Due to privacy considerations, we only release a small portion of the original data. The full dataset will be made available upon acceptance.)