Enhancing levenshtein distance algorithm for assessing behavioral trust

摘要

This paper deals vvith the joint use of a trust evaluation approach and access control mechanisms lor improving security in Web-usage. Trust ('valuation is achieved by means of behavioral analysis, in such way that dilferent access policies can be automatically selected whenever a user behavior is validated. Behavioral analysis uses machine-learning techniques to gain knowledge about user past navigation paths, creating a user signature to be compared against the current behavior of that user. An enhanced version of the Levenshtein Distance algorithm is used to quantify the behavioral analysis (trust calculation), incorporating a way to weigh in the presence of identical subsequences within two distinct navigation paths. This approach is validated through an experimental evaluation which continued that the proposed enhanced Levenshtein Distance technique promotes belter results.