Data Mining and Prediction Algorithms for Takeout Purchase Behavior

In recent years, the food delivery industry has witnessed significant growth, with a rising number of individuals turning to online platforms for their meal orders. Understanding and predicting consumer purchasing behavior in this context is essential for businesses to enhance their services and optimize user satisfaction. This paper presents a study on data mining and prediction algorithms for food delivery purchasing behavior, utilizing deep learning methods. By leveraging deep learning techniques, we analyze user data to uncover patterns and trends in purchasing behavior. The primary aim is to develop accurate prediction models that can forecast future food delivery purchases with high precision. This work explores intelligent prediction of takeaway purchasing behavior through deep learning. Firstly, this paper introduces multi-head self-attention into the scenario of predicting takeaway purchasing behavior and models the dependency relationships between users' historical purchases using positional encoding. The paper utilizes the AUGRU structure to filter the evolutionary path of user interests, obtaining a more accurate representation of user interest vectors. Secondly, this work combines the aforementioned technologies to propose a user takeaway purchasing prediction network for recommendation of takeaway products. Thirdly, comprehensive experiments on the proposed method validate its correctness.