A method for discretization of continuous attributes in network performance measurement

Network performance measurement is an important step to improve network performance. A new method for discretization of continuous attributes is put forward in this paper to overcome the limitation, which the parameters of network are almost continuous attributes.The method is based on an improved algorithm to produce candidate cut points and an algorithm of reduction based on variable precision rough information entropy. With the guarantee of consistency of decision system, the method can reduce the number of cut points, and improve efficiency of reduction. Experiments show that the algorithm yields satisfying reduction results.

[1]  Cai-Yun Chen,et al.  Study on discretization in rough set based on genetic algorithm , 2003, Proceedings of the 2003 International Conference on Machine Learning and Cybernetics (IEEE Cat. No.03EX693).

[2]  Randy Kerber,et al.  ChiMerge: Discretization of Numeric Attributes , 1992, AAAI.

[3]  Song Xiao-xue Rough Sets Theory and Its Application , 2005 .

[4]  Zdzislaw Pawlak,et al.  Rough Set Theory and its Applications to Data Analysis , 1998, Cybern. Syst..

[5]  Robert Susmaga,et al.  Analyzing Discretizations of Continuous Attributes Given a Monotonic Discrimination Function , 1997, Intell. Data Anal..