Linear-Based Recommender Algorithm for Enhancing the Accuracy of Ancillary Service Recommendations in Civil Aviation

With the rapid development of the global aviation service industry, the recommendation and sale of ancillary services play a vital role in enhancing user experience and optimizing resource utilization. However, current recommendation methods often rely on manual rules or simplistic management techniques, showing limited performance in predicting passenger purchasing behavior, which results in low recommendation success rates. Traditional recommendation algorithms in the civil aviation domain are primarily based on conventional machine learning methods. These methods, however, exhibit limitations when dealing with large-scale data and cold-start problems (i.e., lack of interaction data for new users or new services). To address these challenges, this paper proposes a recommendation system based on a linear model to improve recommendation performance in cold-start scenarios and optimize computational efficiency for large-scale data processing. By modeling the relationships between features linearly, the proposed algorithm is specifically designed for scenarios with sparse interaction data, achieving a balance between computational efficiency and recommendation quality. This approach overcomes the high memory requirements of traditional deep learning methods. Experimental results on three public datasets demonstrate that the model significantly improves recommendation accuracy compared to existing methods while showing clear advantages in computational efficiency and resource usage.