Research And Application Of Service Recommendation In Mobile Environment

With the development of wireless network technology and the popularity of mobile phones and other mobile device, mobile device has become a major platform for people to access to information. However, with the rapid development of information technology and the growing information content,"information overload" problem is getting worse. It is a great deal of burden for users, especially for mobile users to access information. Recommend system applied as an effective means to alleviate this problem. But still there are a lot issues that need in-depth and meticulous research to improve the performance of mobile recommendation.We combine content-based recommendation and collaborative filtering recommendation algorithm to alleviate the cold start problem. We use content-based recommendation method when the user behavior data is relatively small or nonexistent. And use collaborative filtering recommendation algorithm when user behavior data is wealthy. At the same time, we can combine the user’s location with the recommend system to recommend. Therefore, we first get the user’s interest keywords, use the keywords to create user interest model. We combine the traditional recommendation algorithm with specific geographical scene on the bases of user interest model to make the recommendation results meet the user’s interest, and within a certain period of time up to a certain extent. So that to improve the intelligence of personalized recommendation system.In addition, taking into account the impact of the way of showing results, we use exhibition system as an example, when recommend users, we highlight e user’s industry information to highlight the user’s concerns. When recommend exhibitions, we also introduced by way of keywords to attract the user’s attention. Furthermore, when we show the recommend results, we can provide recommendation treasons to win the trust of users. As to the user experience, we put the recommend button in the left side bar. The user can click the button to get the recommend result, so that, the system has no impact on the users’normal use.Another way, as for the mobile Internet network is instable, we use third-party frameworks to check the internet connection status when request server. If the network is unstable, we have adopted a middle way cache to solve this problem. The client store data temporarily, when detects the network is normal, and then sent to the server in package.Finally, considering the recommended efficiency problem, we combined offline and online services. The server uses the offline way to build and update user interest model on the bases of the data that come from phone. And make recommendations based on the model. When server detects a change in user interest, it pushes recommended calculation result to the client when the user login the system. The results of simulation experiments show that by setting the users’ location, the way eased the cold start problem, but also to some extent, improve the efficiency and effectiveness of the recommendation system.