| As MIIT of4G licences granted, mobile Internet is bursting with its own another spring.4G network---as the Internet speed, short delay time, traffic synonymous with lower price, not only can realize mobile high-speed data services, and will be better to satisfy the requirements of mobile users’high-speed wireless Internet access. Additional mobile terminal portability, there will be more and more users through a mobile terminal access to the Internet, so the behavior of traditional Internet users will gradually shift to the mobile Internet. This will provide new opportunities for the development of mobile Internet, together with new challenges.In the21st century, not only an information age, but also paying attention to the era of time efficiency. In the time of mobile carrier, the electrical Commerce is popular, who can steal a march on ahead of time, the precision of services and products to communicate effectively to the user, to provide users with personalized business, so that they can maintain its leading role in the same industry and create the biggest benefit. This is an opportunity and a challenge. Then the key technologies for mobile users online behavior research and analyze the user behavior online, understand the characteristics of users surfing the Internet, thus providing high value-added precision of personalized service and high quality products is imperative.First, this paper states the research status of mobile user behavior, key technologies and research methods, and expounds the user behavior analysis of the key technology. Through analyzing the existing methods of the mobile user behavior, we put forward a method of analyzing mobile user behavior by using multidimensional sequence mining algorithm. This method excavates the frequent browsing mode of mobile user, and then considers other dimensions of information, such as user’s gender, age, geography, etc. It can more fully reflect the characteristics of users surfing the Internet and the potential demand of using the Internet, can offer reference for administrators for website optimization and adjustment, and can provide a basis for operators to adjust the marketing strategy and mode, which can offer users with strong effective service and high quality products.Second, researching the current existing sequential pattern algorithm, such as GSP, SPADE, PrefixSpan, PSPM, etc., and multidimensional sequential pattern mining algorithm, such as Seq-Dim, Seq-mdp, and example analysis compares the advantages and disadvantages of them. After analyzing the existing algorithm, we put forward the improved3PC-MDSP algorithm. It is similar to Seq-Dim and Seq-mdp algorithm, which first mine sequence pattern, and then dig out the multidimensional pattern. Only3PC-MDSP uses the improved3PC algorithm to mine sequence pattern, which builds prefix projected location tuples. It can avoid building projection database through constructing the prefix projected position tuples, which scan database by a tuple projection position. And through comparing the different tuples prefix and projection location to avoid the frequent items of repeat scanning, and comparing tuples number and the size of the minimum support, if below the minimum support, don’t keep digging, to avoid the unnecessary waste of time, the frequent pattern mining to improve the efficiency of the algorithm. At last, data generated by IBM data generator, verifies the validity of this algorithm.Finally, designing and implementing a system of mobile user behavior analysis. Introducing the mobile user access to the Internet server side Web-log format, processing methods and data storage method. Through applying the improvement algorithm3PC-MDSP to the system, it can analyze the user behavior online。... |
PDF Full Text Request