Computer Users Activity Analysis Using Recurrence Plot

We can currently observe trend of gathering vast amounts of data about systems and users behavior. Gathered data is analyzed to get insight about behavior of users and then to detect traits of users. Currently many different methods are used to analyze data and there is still no one best method for analyzing different parameters of computer systems. We use non-linear methods to analyze gathered data because we assume that computer system is the non-linear one because of inter-program dependencies in current multi-program, multi-user operating systems. To best describe behavior of entire system we choose number of interrupts per second as analyzed variable because this value shows overall state of activity of programs being run on the system. We show that using recurrence plot can show some similarities in behavior of system, and therefore can be used to detect similarities and differences in users behavior.