A evaluation method for classified security protection of information systems based on neural network and rough set

While carrying out the evaluation of classified security protection of information systems by means of the traditional expert evaluation method, the evaluation results are less accurate and more subjective, and largely depend on the professional ability and working experience of the evaluator. In order to solve this issue, this paper presents an evaluation method for classified security protection of information systems based on rough set and neural network. In this method, firstly, by using fuzzy theory, the quantitative decision table of evaluation indicators can be established with the help of domain experts, and then by using rough set, an attribute reduction algorithm is used to preprocess the sample data in the decision table, finally, these sample data can be used for BP neural network training. The trained neural network is taken as the evaluation model of classified security protection of information systems.