Title :
Classifying User Behavior in Linux Sessions
Author :
Fox, Richard ; Davidson, Eric
Abstract :
Classification is an integral part of many types of problem solving such as diagnosis, understanding, and as a component in prediction and design. Here, hierarchical classification is used to determine a user´s behavior in a Linux environment as the user enters commands via a command-line prompt. Once successfully classified, one might be able to either predict what the user is trying to accomplish and help assist the user, or perform error correction when the user types in a command that is either syntactically incorrect or contains incorrect parameters (e.g., misspelled filename). This paper reports on the first part of this project, using hierarchical classification to correctly identify what a user is trying to accomplish, and briefly discuss future work on error correction and prediction.
Keywords :
Linux; human factors; pattern classification; program debugging; program diagnostics; Linux; command-line prompt; error correction; error prediction; hierarchical classification; problem solving; user behavior; Artificial intelligence; Bayesian methods; Classification tree analysis; Computer science; Data mining; Error correction; Genetic algorithms; Linux; Neural networks; Problem-solving; AI Classification; correction; prediction;
Conference_Titel :
Tools with Artificial Intelligence, 2008. ICTAI '08. 20th IEEE International Conference on
Conference_Location :
Dayton, OH
Print_ISBN :
978-0-7695-3440-4
DOI :
10.1109/ICTAI.2008.19