Research And Implementation Of Personalized Recommendation System Based On Tourist Behavior

With the hot travel market and the rapid growth of travel information,online travel applications have become an important industry application for the Internet.Since many users find it difficult to find their favorite attractions in time,the online travel recommendation system emerges as the times require,providing users with attractions that may be of interest to and helping users complete their travel plans.Many existing travel recommendation systems focus on the discovery of user's interest attractions,and believe that targeted recommendations for user's interests can meet the needs of users.However,they do not have a deep understanding of the travel industry and users'personal characteristics,so it is difficult to achieve the effect that satisfies the user.The personalized recommendation system for travel has become the focus and direction of research.This paper designs a personalized attraction recommendation system for users based on Internet big data.The main work of this paper is as follows:(1)Obtain large-scale data of online travel websites through distributed crawlers,including tourist attractions and user ratings.By analyzing user data,it is found that different curiosity of users leads to different choices of their attractions.Therefore,a personalized recommendation scheme for tourists based on user curiosity was designed.(2)In order to better describe the curiosity of users when selecting attractions,a approach of attractions classification suitabele curiosity is proposed.Keywords are extracted according to the TF-IDF value after the word segmentation of the online attraction introduction text,and the keywords are converted into a word vector to train the attraction classifier.(3)A personalized recommendation algorithm based on the curiosity of tourism users is proposed.The novelty calculation formula is used to transform the user's historical behavior into a novelty distribution map representing the curiosity of the user.The novelty represents the familiarity of the user with the attraction.The algorithm is divided into two parts.The first is to calculate the novelty selection map of the user based on the user's historical access record using the novelty formula,and the second is to obtain the user personalized recommendation scores of the novelty from the distribution map.Combining the result of the collaborative filtering algorithm and the user personalized reconmmendation scores to realizes personalized recommendation of the attraction.It has been verified by experiments that the performance of the proposed recommendation algorithm has been improved.(4)Design and implement a personalized recommendation system for travel.The system which is based on the travel personalized recommendation algorithm is mainly composed of data acquisition layer,data processing layer,model calculation layer and business application layer.The data acquisition layer completes the collection of the original data,the data processing layer completes the specification work of the data,the model calculation layer completes the process of the attraction classification and the personalized recommendation algorithm,and the business application layer provides the user's service interface.