Research And Implementation Of Tourism Recommender System In Mobile Environment

At present,with the development of mobile internet and tourist industry,each tourist business website,such as XieCheng,YiLong,QuNa and so on,developed their respective mobile Apps in succession while there is a defection in most of these Apps that lacking of a function which could provide personalized recommendations on sites,restaurants,hotels etc.Traditional tourist recommendations only take user basic characters and historical behaviors into account,which hardly consider circumstances in the process of the tour.Nevertheless,tourists' preferences may change with their changing circumstances,even tourists in the same circumstances with different personalities.So,in this paper we focus on a deep research on personalized recommendation in the field of tourism considering both factors of context and behavioral preferences.The main work is as follows:To analyze and summarize the key context factors that may influence users' decision during the tour including time,location,weather and so on,by means of investigation and study.To read relevant information,to learn about the obtain method and expression method of tourist context and try to retrieve these context from mobile Apps during the tour.To analyze the context characters and preferences and develop a recommendation model based on context awareness and collaborative filtering recommendation.To calculate the similarity of context and historical context,pre-filter the context that irrelevant to the current context.To calculate the similarity of object user and users in similar contexts,find out the group of users who have similar preferences to the object user,calculate the grading matrix of the group,predict new item of the object user and return the Top-N recommendations to the object user.Finally,we implemented the App on the platform of Android mobile devices and applied the tourist recommendation model in mobile context that we proposed in the system in order to fulfil the target of personalized recommendation.We established a test set by means ofreleasing the App and collecting data.We also tested the performance of recommendation with various evaluate method and refined the model.Results,analyses and conclusion we summarized at last.