Transductive quantitative calculation approach of network abnormality index

Abstract

For the problem of network abnormality index calculation,a transductive quantitative calculation approach named QCDP was proposed based on quantitative characteristics index,composition characteristics index,distribution characteristics index and pattern characteristics index.Seven experiments using real network traces were made to validate the effectiveness of QCDP.Theoretical analysis and experimental results show that,compared with the traditional transductive method based on traffic,the QCDP can more effectively reflect the macro situation of typical network security incidents; compared with the inductive methods,the QCDP has better objectivity,instantaneity and practicability.

Keywords

• network security
• macro situation
• abnormality
• index
• transduction