Feature Selection Algorithm Based on Cloud Model and Rough Set
Huang Qiao-yu · Journal of Dongguan University of Technology · 2015
The network security log has a lot of attributes,and some of them have no effects or few effects on network situation analysis,which may cause a waste of time. This paper presents an algorithm of feature selection based on cloud model and rough set. Cloud model is applied to sort the importance of log attributes in intrusion detection system,the result of which can be adjusted by the opinion of experts,guaranteeing the attribute order to accord with the actual distribution of data and satisfy the demand of attribute preference,making the attribute reduction of rough set theory to reduce the dimension of data vector,automatically selecting the effective feature subset according to the given attribute order. The results show that the algorithm is feasible and efficient.