Research On Social Network Based Travel Route Recommendation Algorithm

With the rapid development of economic and improvement of people's living standards, self-traveling gradually becomes the first choice for people's traveling tourism. However, with the rapid development of multimedia and network technology, the amount of tourism resources increases larger and larger. Thus, people have to spend lots of time and effort before traveling to draw up their own travel plans. Although there exists many travel systems in the network, it is difficult for them to meet the user's personal needs. From intelligent tourism point of view, we have a deep study on travel route recommendation in order to meet the user's personal needs. The main contribution of this paper includes three parts: the study of user's interest model, the travel route algorithm based on user interests and the web system of travel route recommendation. The author's detail contributions are outlined as follows:1) user interest model based on user visiting history are proposed. The entire process is listed as below: firstly, using Bo W model to obtain the whole words frequency, and then constructing the dictionary through stopwords,tagger and Word Net filtering, simultaneously using TF-IDF to get the landmark's vector representation, the landmark's feature is obtained by mapping the vector representation to the dictionary, and then calculating the user's feature based on user's historic attractions and landmark's features, finally constructing the user's interest model through the similarity measure between user feature and landmark feature.2) a travel route recommendation based on user interest is proposed. Specifically, the proposed method has three parts: at first, using markov model to build the database of landmark transfer probability through the total downloaded Flickr images; then, building the user's interest model; finally, combining the markov model and user's interest model. The experimental results show that the proposed method has good performence.3) a travel route recommendation system based on Struts2, Hibernate, My SQL is developed. First, analyzing the design of the travel route recommendation system from the following four aspects: economic feasibility, legal feasibility, technical feasibility and management feasibility. And then, introducing the related technology roughly. At last, bringing the implementation of the system in detail, including the design of the database system, front page and background functions. The travel route recommendation system includes the following modules: user registration and login, landmark data show(including attractions microblog, travelogue, pictures, etc.) as well as the travel route recommendation. As the core module of the system, travel route recommendation works as follows: the user initiates a request through the front-end submission form data(including the current location attractions, the user's historical landmark records) to the background; and then, the background calculates the recommended results through the proposed method and returns them to the front page; by clicking on the recommended results, the result will be visualized through the Google Map.