Research And Implementation Of Recommendation System Based On User Preference And Product Tags

Throughout the entire airline business in China, it keeps a steady and continuous increase in recent years. As far as the airline network is concerned, its scale increases by 10% yearly. With the development of airline route, there are increasing products based on airline routes. The passengers have more choices to go shopping when they travel. The passengers can choose the products that they bought previously as well as the popular ones. When selecting, they have various needs. Some passengers have enough time and they tend to hunt for new products or tourism products. Others who are busy because of work or other reasons,will not pay attention to the information of tourism products. Instead,they will consider the convenience when they are purchasing. However,in the present purchasing system, all of the available tourism products are offered and it is the clients who make choices. In this way, it is not only a waste of time but also hard to meet their needs.Therefore,the recommendation algorithm has a great research value to list the tourism products according to the clients' specific requirements.The thesis is first in the prototype of recommendation algorithm based on tags and Hadoop+HBase. The original algorithm is improved and the recommendation based on user preferences and product tags is also given.According to the user's personal information, purchase history and browsing history, the user tags are given from various aspects, such as the user's preferences and the history of the user's purchase. And it will recommend a trip or product package that are suitable to the user's characteristics. The recommendation algorithm can also be applied in the airline route tourism website.The thesis in the core of recommendation algorithm designs and accomplishes the simulation system to recommend the tourism products.When the clients log in,they can check relevant recommendation list and the clients can make choices of their favorite products. The system is based on the B/S design mode. The background system deals with the data changes and inputs the results into the HBase. The foreground checks the required information directly and shows them to the clients. In this way, it can not only deal with the changes of tourism product information but also can satisfy the real-time requirements when they are inquiring. At last, the thesis makes use of the present data to simulate huge amount of verified recommendation results. The results show that the recommendations based on client preferences and product tags achieve a great effect in the product recommendations of tourism websites.