User Modeling: Through Statistical Analysis and an EvolvingClassifier

摘要

Knowledge about computer users is very beneficialfor assisting them, predicting their future actions or detectingmasqueraders. In this paper, an approach for creating andrecognizing automatically the behavior profile of a computeruser is combined with an evolving method to keep up to datethe created profiles. The behavior of a computer is representedin this research as the sequence of commands s/he types during aperiod of time. This sequence is treated using statistical methodsin, order to create the corresponding user profile. However, as auser profile is usually not fixed but rather it changes and evolves,we propose a user profile classifier based on Evolving Systems.This paper describes briefly the model creation method and theevolving classifier, which are compared with well establishedoff-line and on-line classifiers.