| With the rapid development of technology and the increasingly mobilecommunication technology, more and more people in today’s society with mobilecommunication devices to communicate. This will inevitably give mobile operators tobring some burden of communication equipment.Especially in the important activitiesor hold some major festivals, there will be a violent traffic growth.In order to makemobile communication network operators and equipment stable and safe operation, aswell as to provide a better mobile call quality service, predict in advance of the trafficis very important. Although research on the traffic prediction has years of history, andthere are already some methods whichis applied tothe actual, however, these studiesare based on theoretical models and did not form a specific system to be dedicated tothe traffic prediction. In order to provide an operability better forecasting system forXinjiang Mobile, it is very essential todevelop a kind of practical traffic predictionsoftware.This paper describes the design of a traffic prediction system based on MATLAB,through of Xinjiang traffic date analysis and processing, using MATLAB powerfuldata processing capability to achieve operational analysis of the data, after analyzingtraffic data2004-2012, SMS data and user data, and methods to predict the directionof a large number of domestic and international research, models, developed a PSOsupport vector machine method, genetic algorithm optimization support vectormachine method, hidden Markov forecasting method, gray prediction method, Elmanneural network prediction method of five. Finally, MATLAB/GUI interface to thesystem for the preparation and realization of the whole system integration. And use itfor the2013mid-autumn festival trafficprediction and error analysis, forecastingresults are generally able to achieve the desired requirements.The main contents of this system:(1) this system is to achieve the major holidays of each state’s forecast, and be able to predict the outcome of the icon.(2)Development of user-friendly, graphical interface, with a clear prediction interfaceand easy to understand, easy to use it, and it’s predict result is very reliable.This system is currently available for long-term forecasting, not only has a longlife cycle, it is easy to maintain it. |
