Fighting malware and spam

A novel approach to improve scanners using fuzzy logic

Amit Saxena G.G. University, Bilaspur, India
Anshuman Dwivedi G.G. University, Bilaspur, India

Anti-virus programs or scanners are one of the common ways to protect our precious data from viruses. In conventional techniques a test set of viruses is formed for evaluating the performance of a scanner. This test does not take any care of the characteristics of the viruses included for test. Further, this random selection may result in creation of test set lacking representation of severe viruses for testing. Secondly, the existing techniques for comparison of scanners for certification require how many viruses are missed by a scanner, irrespective of kind of damage caused by viruses. In both cases, the qualitative behaviour of viruses is not considered in conventional techniques.

In this paper a Fuzzified Scanner Comparison Technique (FSCT) is proposed which determines the ranks of the viruses on the basis of three parameters and, based on that, suggests formation of a ranking-based test set. The same scheme can be applied for comparing scanners by marks scored by them during testing. The technique is simulated on sample data to justify its potential to be adopted for selection of a scanner to recommend for certification by various virus test centers.