Modifying Erlang B table based upon data mining

During the process of mobile network planning and optimization, capacity predication is one of the important contents. Accurate capacity results will not only improve the running quality of the network, but also will improve the satisfaction of the users. Traditionally, Erlang B formula was used to forecast the capacity of network. The forecasted results were quite suitable for the wire network. When the same method was used in wireless voice traffic predication, the results are inaccurate for the difference between wire and wireless channels. To precisely forecast the traffic capacity of a wireless network, the Elang B formula should be modified to fit the characteristics of wireless systems. In this paper, using data cleansing based on clustering algorithm, the algorithms and methods of data mining, the modified Erlang B table including system design, process flow was implemented. The field test was given to show the comparison between theoretic and modified results.