Privacy-preserving dish-recommendation for food nutrition through edging computing

The catering industry is a humungous service-based industry, which includes food nutrition and security, transporting, tourism, and similar services. Employing the new techniques, such as machine learning and deep learning, competing firms can improve their strategies and deliver better services with lower price. In this work, we investigate a large number of historical dining data that reflect customer spending habits and taste preferences through terminal devices at the edge of the network, which can be used for customers to recommend dishes when ordering. With the popularity of cloud computing today, it is also prone to some shortcomings when the amount of computing is too large, such as poor real-time performance, high server pressure, data security risks, and so on. Especially, the leakage of user privacy often has incalculable consequences. Based on this, this article proposes a hot pot dish recommendation algorithm based on data mining and food nutrition for edge devices, which can effectively protect user privacy. The experimental data in this article is derived from the massive consumption data of customers in China's large hot pot enterprises. First, clean the raw data and mask the user privacy to get feature data unrelated with user privacy that can be used for dish recommendation. On the basis of this, combined with data mining methods, diet, and other indicators, the results of final recommended dish can respond to both user preferences and nutrition arrange. The experimental results show that this dish-recommended method is more reasonable and healthier.