A Research On Destination Recommendation System Based On Collaborative Filtering Method

With the continuous development of information technology, consumptionpatterns of tourists have great changed.The trend of relying to information isstrengthening, the application of network resources is extending, and informationrecommendation technology is one representative of them. Recommendationtechnology is to help users screening unnecessary information, making the users getthe most effective information, reducing the cost of time and funds for collecting,collating and analyzing data. The recommendation technologies mainly includecollaborative filtering method and content analysis method. In the research,considering the collaborative filtering method as the main line of research, makingfull use of the important soul of content analysis in the tourism resource development.Finally, the research have been forming a model that collaborative filtering method isthe principal line and content analysis method is an important aid.In the research, strengthening the introduction and distinguish of theoreticalcontent about recommendation technology, collaborative filtering method, contentanalysis, DMS system and tourism destination marketing. The theoretical basis andrelevance of the theory of the research have been rationalized. On the basis, basicelements and operation of tourism destination recommendation system have beenfurther studied. Firstly, discussing the basic elements of the recommendation system,including destination eigenvectors, tourists experience vector and touristseigenvectors. Destination eigenvectors mainly include:(1) the study of the charactersof destination information input, finally tourism destination management departmentis defined as information input body and according to the national institute ofstandards to divide the destination resources;(2) The study of form templates ofdestination information. Input includes destination area, destination name, rating,phone number, tourism management department, URL, travel characteristics anddestination development time. It more specifically shows the destination characteristics and provides practical convenience for tourists.With regard to tourists experience vector, the research takes close attention to thetravel experience of tourists, by calculating the similarity of the similar preferences oftourists, according to the higher similarity of trivial experience, suggesting a newtourist destination to the target tourists. In consideration of the following three cases,designing input form template:(1) The tourists who have inputted trivialinformation;(2) the new tourists who have not inputted information;(3) the newdestination. The research has designed the different input templates for the threecases.Concerning Tourists eigenvectors, the research sets up eigenvectors Ft of thetourist t, and by the tourists experience matrix H and destination feature matrix Ecalculating automatically. The tourists visit the more destinations, Destinationcharacterized the higher concentration, and the more accurately reflect the personalpreferences of the tourists.By studying the basic elements of the recommendation system and finallydiscussing the specific calculation method of the tourist destinations recommend,finally bringing forward to the improved algorithm of collaborative filtering based onitem classification. Firstly, making use of utilization projects classified informationand adopt the clustering technology to forecast the score for ungraded items.And then through calculating similarity within-class to obtain the target user’snearest neighbors. Finally, form the recommend. In the end of the article, regarding tothe problems in the course of the study, making the following recommendation todevelop the applied research of the system.(1)With the constant development of collaborative filtering recommendationmethod in practice, promoting the system to reflect the change time and trends of thetourists′interest timely and accurately.By grasping the different time stages of thetourists′interest, and the continuous phase of the clustering analysis to Promote thegrasp of tourists similarity.(2)Promoting the construction of indicators reflecting tourists′interest.Designing the indicators that tourists interest in, inputting the information by themselves, drawing support from software analysis to grasp the tourists′differentstages Characteristics in advance.