The Research And Application Of Food Recommendation Method Of Campus Based On Context-aware

The rapid development of information technology in today's society,the application of personalized recommendation technology to the catering industry is not only beneficial to technological innovation,but also beneficial to the development of the catering industry.Compared with the traditional project recommendation,food and beverage recommendation is more dependent on the user's own and physical environment factors,and dining options in different environments are also quite different.Therefore,the current online food and beverage recommendation is based solely on the user's past behavioral data to generate the relevant food recommendation model to solve the food recommendation situation.Sensitive issues seem powerless.Therefore,based on contextual awareness technology,content-based recommendation algorithm and user-based collaborative filtering algorithm as the research premise,from the perspective of practical application of campus catering,the paper proposes a solution to the problems of “cold start” and “data sparsity” existing in traditional recommendation systems.A school dining recommendation model that combines the current situation and interest factors of students and detailed design and implementation of a prototype recommendation system are used to validate the proposed model.The main research contents of this paper are:(1)Define and explain the situational awareness framework,and establish the student's diet context ontology model from the perspective of situation definition,expression and modeling;after determining the scope of knowledge in the catering field and acquiring domain knowledge,use the form of the five-tuple A catering knowledge ontology model was established and perfected from the top down;the situational inference rules were divided into context deduction rules and recommendation optimization rules according to different perception stages,and were specifically designed.(2)Starting from the problems existing in the traditional recommendation algorithm,it proposes to collect student behavior data to draw the student interest feature vector,and after the similarity is calculated,the user is recommended to be based on the existing interest in the food and beverage project;then an improved collaborative filtering algorithm is proposed to the user.The recommendation is based on potential interest catering items.The principle is to predict the scores of catering items that can be filled by users based on the user interest eigenvectors to obtain more accurate recommendation results.Finally,the algorithm based on situation recommendation and recommendation based on user interests is recommended.The models were merged and the process described.(3)The client and server of the prototype system for food and beverage recommendation were designed in detail,and the feasibility of the proposed method was verified through experiments.