Design of the Tourism-information-service-oriented Collaborative Filtering Recommendation-CodeShoppy

Design of the Tourism Personalized recommendation technology is one key application in modern Electronic commerce field with optimistic prospect. As the urgency of the use of recommendation technology in tourism industry, the authors try to design a collaborative filtering recommendation algorithm integrating with nearest neighbor recommendation and cluster analysis referring to national criteria “Classification, Investigation and Evaluation of Tourism Resource “on the base of existing collaborative filtering recommendation technology. Theoretical mechanism and realization method of this improved collaborative filtering recommendation algorithm will be also discussed. Demand of user is often ambiguous and indefinite, it may contain potential demand for some product without specific target. And if we could recommend this product to user, it is to turn user’s potential demand into real demand. Traditional user-search service model can not meet real and efficient demand. In America artificial intelligent conference in March of 1995, Robert Armstrong et al. produced a personalized navigation system named as “Web Watcher” with recommendation technology as the kernel [1]. Later on large E-C system such as Amazon, CDNOW, eBay, MovieFinder, Reel have started to adopt recommendation technology. Some Chinese on-line shopping websites such as Taobao.com, Sina.com, Netease.com, sohu.com and DangDang.com have produced their own E-C recommendation systems. Earlier personalized recommendation is regarded algorithm as the core. Konston suggested collaborative filtering recommendation technology to produce recommendation outcome according to the similarity among users and items which enjoys relatively better real-time and recommendation quality.(read more)

Then the introduction of new knowledge such as information acquisition and artificial intelligence and fuzzy recommendation expanded the thoughts of recommendation algorithm. Zhang & Guo designed a collaborative filtering Recommendation algorithm based on item evaluation predication, it is a new similarity measurement method to calculate nearest neighbor of target user [4].As for the problem of extensibility in personalized recommendation system application, Li et al. put forward mass-customization personalized recommendation algorithm[5], Deng et al. summarized item-based collaborative filtering algorithm which could effectively resolve the problem of cold start in recommendation system to improve the quality in recommendation system efficiently [6]. Lin et al. mentioned about information recommendation mechanism based on contents and collaborative model [7], and there appeared resource collaborative filtering technology based on semantic similarity [8, 9].As one of the most widely used and successful technology in personalized recommendation system, collaborative filtering technology includes user- based collaborative filtering and item-based collaborative filtering [10]. The former is to find neighbor user of target user according to similarity among users, and then to propose the recommendation for target user referring to past information of neighbor user. The later will analyze the similarity among items according with target user’s bought favorite item and recommend similar item for target user [11]. Authors think these two methods can be integrated to realize recommendation tailoring with industry characteristics of tourism information service. First, people can cluster the items with same features, and then to cluster them by user’s interest similarity to find the nearest neighbor of target user, and select the product of higher nearest neighbor score with shared preference as target item.