A Recommendation Algorithm Incorporating Adaptive Gating Mechanisms and Knowledge Graph Enhancements

Aiming at the problem that existing knowledge graph-based recommendation algorithms do not fully utilize the interaction information between users and items, this paper proposes a recommendation algorithm called Gated Adaptive and Knowledge-Enhanced Recommendation (GAKR) that integrates an adaptive gating mechanism with knowledge graph enhancement. First, GAKR uses a Multilayer Perceptron (MLP) in the recommendation module to process the user's initial feature vector and extract potentially compressible features. Then, for the item feature vector, the model dynamically adjusts the weights of Collaborative Filtering (CF) and Knowledge Graph (KG) through the gating mechanism, generating a more expressive item representation. The gating mechanism calculates the gating value through the sigmoid function, allowing user's behavioral characteristics in different scenarios to more accurately impact item recommendation. Next, in the Knowledge Graph Embedding (KGE) module, an embedding interaction mechanism is designed, which updates the entity embeddings by using Mobius addition and mapping, capturing higher-order interaction information between entities and relations. To prevent overfitting, GAKR employs L2 regularization during the optimization process, which enhances the model's generalization ability when handling complex interactions. Finally, a normalized inner product operation is used as an evaluation function to predict the user's preferences for items. The experimental results on three public datasets from different domains – MovieLens-1M, Book-Crossing and Last.FM – show that the GAKR model outperforms other benchmark models in terms of the evaluation metrics such as AUC, F1-score and recall rate in the CTR (Click Through Rate) and Top-K recommendation scenarios.