Telecommunications Traffic Olap Analysis And Neural Network Prediction Model

Having applied computers for many years, people have accumulated a large number of data. On-Line Analytical Processing (OLAP) and Data Mining provide effective tools for data analysis. With the development of the telecommunication industry and expansion of telecommunication networks, telecommunication corporations face the challenge. It is getting more and more important for telecommunication corporations to analyze and predict Telecommunication Traffic efficiently to help their business. However, traditional traffic analysis methods pay attention to current status of network and deal with few of historical or future data. At present, the telecommunication companied are still pressed for traffic forecast algorithm and tool in operation network system.This paper researches the techniques and implementation of telecommunication switching network traffic analysis based on OLAP. It is motivated by the requirement of switching network maintenance and follows telecommunication network technical standard. My work includes:(1) Investigating the characteristics of original traffic data of exchanger in telecommunication network, the requirements of analysis and forecast, as well as KPI (key parameter indication) in network operation and maintenance.(2) Creating a model for telecommunications traffic analysis. The traffic data is brought forth by way of the many-dimensions, and the dimension degree of data is defined.(3) Developing a tool, which provides various query analyses, and makes various personnel understand directly information that is contained in the data? And according to the particular requirement, it realizes the nimble query. The users can analyze data neatly by way of the unified WEB Browser interface, need not consider the operating system.(4)Having accomplished the forecasting model based on the neural network using MATLAB tool, which is a new thought and algorithm to traffic forecast of telecommunications enterprise. The forecasting is tested by using the time queue. The historical data of exchanger is regarded the input of neural network using BP algorithm. When the experiment is finished, we get the law of the concerned traffic changes. The length of data window, which is betaken to calculate the coming mensal traffic value, is determined by the traffic character of different districts. In order to better understand the effect of neural network forecasting, the paper compares the forecasting result using the way of the neural network with other models under the identical circumstances.At present, the technology of On-Line Analytical Processing and Data Mining is used for the competition of the market and data analysis. And along with the constantly enlargement and complication of telecommunications network, their application in the operating maintenance of telegraphic network will be great.