Design And Implementation Of Healthy Diet Recommendation System Based On Multiple Features

In recent years,people have paid increasing attention to a healthy life,and a healthy diet is an important factor in ensuring good health.However,individuals often lack accurate judgments on how to effectively arrange a daily diet.The implementation of a diet recommendation system can help people design and choose a reasonable diet plan.Therefore,this thesis designs and implements a multi-feature-based healthy diet recommendation system based on the user's diet preferences,contextual information and nutritional balance.The main contents of this thesis are as follows:Firstly,this thesis summarizes the current research status of recommendation systems and diet recommendation services at home and abroad,fully investigates the diet software and websites on the market,clarifies the software architecture of this thesis and the software technical solutions.The implementation of the healthy diet recommendation software is divided into two parts: the user side and the administrator side.Secondly,this thesis improves the collaborative filtering algorithm and applies it to the diet recommendation system with the ontology-based contextual reasoning method and the multi-objective particle swarm algorithm.First,the improved collaborative filtering algorithm is used to infer the users' taste preferences and calculate the users' predicted scores for potential dishes.Based on term frequency-inverse document frequency,the algorithm improves the accuracy of similarity calculation by introducing user influence factors and dish influence factors and corrects the influence of time on users' interest by introducing different time weighting functions.Second,this thesis uses the ontology-based contextual reasoning method to infer the dish set suitable for the current situation,and matches the result with the dish set that meets the taste preference,and sorts and recommends the dishes according to the predicted scores of the dishes,which effectively improves the rationality of the recommended dishes.Finally,according to the main ingredients contained in the dishes selected by the users and the nutritional elements required by the users,the multi-objective particle swarm algorithm is used to calculate the consumption of food for each meal,so as to arrange a nutritionally balanced meal plan for the users.Thirdly,this thesis carries out the software development and testing of the healthy diet recommendation system.First,the development tools and testing tools for the mobile and server are determined,and then the detailed design of each module is completed,and the expected functions are realized according to the design process.Finally,the functions and performance of the software system are tested separately after the software development is completed.The test results show that the system meets engineering application standards and achieves the expected goals.