Research on Personalized Recommendation Technology Based on Collaborative Filtering

Personalized recommendation system is hot territory in the research and application of Internet enterprises in recent years. In this paper, based on the main idea of collaborative filtering, the traditional collaborative filtering algorithm is improved, including data model and algorithm process, considering the actual data, calculation consumption and effect of hotel personalized recommendation. Based on the improved data model, when calculating the recommendation results, the similarity between users and items is calculated directly, which greatly reduces the computational pressure. Finally, considering the different effects of different data, two sub-processes of candidate set fusion and reordering are added to further improve the final recommendation effect. The computations are divided into offline and online computations to maximize performance and fast response algorithms.