Research On Recommendation System For Evaluating Scheme Of Network Security Level Protection

With the promulgation and implementation of the network security law,the requirements for the depth and breadth of the assessment of the level protection of network security subjects are getting higher and higher,and the assessment is getting more and more difficult.An excellent evaluation scheme can not only select targeted evaluation indexes and methods according to the characteristics of the evaluation objects,but also control the difficulty and workload of the evaluation work,so as to find out more safety problems in the evaluation system.The quality of the evaluation scheme often varies greatly depending on the professional of the evaluation personnel.There are many requirements for assessment standards.With the rapid increase of level assessment business,the number of assessment personnel with rich assessment experience is limited,which leads to the lag in the compilation of assessment schemes and the decline in quality.Therefore,it is of great significance to study the automatic generation of computer-aided level protection evaluation scheme.Evaluators compile an evaluation scheme based on the characteristics of the system being evaluated and combined with the requirements of level protection.The difficulties in the process of automating the preparation of the evaluation scheme are:(1)How to select the evaluation indexes that meet the system under evaluation from the huge evaluation index library;(2)How to screen the targeted evaluation methods for the measured objects;(3)How to reduce the number of evaluation indexes in the evaluation scheme.In view of the above problems and the lack of tools for computer-aided generation of evaluation scheme,this paper studies the recommendation system of level protection evaluation scheme from three aspects: evaluation index system,assessment program recommendation(APR),design and implementation of evaluation scheme recommendation system;(1)The evaluation index system needed by the system is established;(2)The evaluation scheme recommendation model is designed,including the design of the similarity calculation method of the evaluated system,the sparsity problem of the scoring data is solved by using the similarity between the evaluation indexes,and the evaluation index score of the evaluated system is calculated by using the scoring prediction method;(3)On the basis of APR model,the recommendation system of network security level protection evaluation scheme is programmed.Due to the high sensitivity of level protection evaluation data,according to the evaluation points,the data set with high similarity in this field is selected as the transfer learning data set.This paper uses the control variable method to set up multiple groups of experiments,optimize the parameters of the APR model,analyze the impact of sparse scoring data on the scoring prediction results,and analyze the evaluation scheme recommendation effect of the APR model.Experiments show that the APR model can find the problems in the system to be evaluated,recommend the appropriate evaluation index to write the evaluation scheme,and the APR model is feasible and effective in intelligently writing the evaluation scheme.