The Research And Application Of A Preference-based Personalized Food Recommendation Algorithm

In recent years,with the rapid development of technology and increasing growth of in-formation,information overload has become a serious problem that disturbs everyone.How to effectively help users to obtain and filter information has become a hot research issue and gains a lot of attention from many scholars,therefore personalized recommendation technolo-gy has been proposed,and it is recognized as a tool that can effectively alleviate information overload.Personalized recommendation is a technology that can help users quickly find useful information by digging the binary relationship between the user and the item.The personalized recommendation does not require the user to provide specific requirements,but rather to ac-tively recommend them information that satisfies their demand and hobbies by analyzing user’s historical behavior.Currently,recommended technology has been widely used in a variety of large-scale multimedia and e-commerce sites,but few are for personalized dish recommenda-tions.In this paper,we improved the traditional user-based collaborative filtering algorithm,and proposed a personalized food recommendation algorithm.Firstly,this paper discusses the concept of personalized recommendation system,intro-duces the principle of content-based recommendation algorithm,collaborative filtering recom-mendation and hybrid recommendation algorithms in details,and analyzes the difficulties that the recommendation technology is now facing.Then,this paper studies the construction of traditional user model in recommender system.Aiming at the difficult feedback,the cold start and the alias of the dishes,the paper proposes a user preference-aware model based on food hierarchy.The model divides the dishes in the restaurant by category,and then constructs cate-gory weight tree to realize the user preference modeling.In order to take full advantage of the expert user’s knowledge,it utilizes an iterative learning model to discover the expert users of each category in the restaurant,which helps to reduce the computational complexity of the user similarity and improve the recommendation efficiency.Secondly,this paper designs a preference-aware selection algorithm to select expert users and candidate dishes according to the distribution of user preference.The results of the s-election are used to predict the user’s ratings on the dishes through the improved similarity formula.Additionally,as the influence of time effect and user interest change in the traditional recommendation algorithms are rarely considered,the time factors are introduced to improve preference-aware selection algorithm and dish rating prediction,and it can successfully recom-mend users dishes in the real scene.Finally,in order to evaluate the validity of proposed algorithm in this paper,we carried out experiments on the data set collected from the cooperative restaurants,and made an anal-ysis of the experiment results.The results show that the algorithm proposed in this paper can improve the precision and recall rate of personalized food recommender system and improve the recommendation quality.