The Research And Implementation Of Personalized Recommendation System Of Dishes

With the arrival of the era of big data, large amount of data be created every second, in result that people often submerged in the sea of data and can't make their choices quickly and accurately faced with the massive information. Against this background personalized recommendation system emerges. The development of food and beverage industry and the mutual penetration throughout the food culture of every where have met the requirements of the dietary diversity of people. But with the substantial growth in the number of dishes and the emergence of new products, Diners have encounter the same problem of choice difficulty when faced with a large number of dishes especially unfamiliar dishes.Personalized recommendation system also applicable to the dishes recommendation. The system personally compute and find the users' interests by studying and learning the users' behavior habits to give the users some recommendation that match their interests.This thesis have applied the personalized recommendation system in dishes ordering after analysing the datas about dishes, and will complete the following works:1) Study the attributes of dishes, and choose the recommendation mechainism based on association rules to study and learn associated relationship between the dinners and the dishes. The system use FP-Growth algorithm to discover the potential association between the dishes from a lot of historical orders and put the association relationship into a set as association rules. Then combine the dishes that the diner had choosed and the frequent patterns to give the dinner recommendation that most likely to meet the taste of this diner. 2) Using the Feedback information that the diner whether to accept the recommendation to improve the algorithm making more accurate recommendation. 3) Analyse the influence by the important parameters in algorithm tochoose the appropriate parameter values that make the user experience be the best. 4) Completed the construction of the recommendation system: design the appropriate data model and the communication protocol, complete coding the server and the client program, and do some custom modification on the source code of the operating system running on the client hardware to make it meet the needs of users.