| With the development of mobile Internet and mobile terminal, the range of information which a terminal-user pays attention to is more extensive and the content of information demand and the ways of obtaining information have nothing in common. Users are increasingly dependent on networks, personalized needs, user experience researching and user perception analysis have become a hot spot. In communication industry, user’s consumption data occupies more and more storage space, these massive data contains a lot of very useful commercial information. If we can mine them, it will be a great treasure. In the face of fierce competition in the mobile communications market, customers are the base of operators, which requires mobile operators rapidly response to a variety of business needs of the user under the premise of ensuring the network quality of service, thereby enhancing customer satisfaction degrees.Nowadays, in the telecommunications industry, common user behavioral analysis methods are as follow:the behavior of the mobile value-added business customers, the network effect theory-based mobile communications in consumer behavior analysis and the travel pattern recognition mobile location-based data users. These are from the operator’s point to analyze user behavior. If communication services of operators are affected, users can not use those normal services, eventually causing customer complaints. How to analyze the behavior of the crowd of those complaints and improve the quality of communication to keep their customers needs should be thinking deeply. The main work is based on those phenomenon mentioned above. Next, we expand the following research:1. Making an analysis for the current wireless communications research background and significance in domestic and international; Have researched the development and current situation of clustering algorithm both at home and abroad; Analyzes the commonly used clustering algorithms and outlier mining technology and their applications, and further analyzes the clustering algorithm in mobile network applications.2. Common wireless terminal positioning methods were studied; Then this paper mainly studied regional signal fingerprint matching positioning technology and mobile terminal positioning algorithm based on base stations; Discussed factors in the signal propagation process which affected the accuracy, and made the corresponding processing; And then altered the two positioning algorithms; Applying clustering algorithm and entropy theory to the fingerprint matching algorithm. And after the modification with multi-base stations positioning method, the accuracy is more reliable.3. This article has analyzed three methods of positioning, in combination with the relationship between user behavior and the method associated with location, by using the fuzzy clustering algorithm, user behavior analysis model is established.4. For the dataset, taking into account the large amount of data, specifications inconsistent etc. we made a preprocessing operation; In the process of calculation, it also used an efficient method of quick sort.5. We have carried out some experiments about the positioning algorithms and user behavior analysis mentioned in this article. Results show that the proposed positioning algorithms have higher positioning accuracy, so we made a detail analysis of the results.Finally, we made a summary of the research and design work, and also pointed out the direction of the next step of work and ideas in the mobile positioning method of wireless networks and user’s behavior analysis, and also described outlier mining of user behavior analysis in the mobile network. |
PDF Full Text Request