Symptom Analysis of a Web Server Log

Problem identification is an area of research from Autonomic Computing. This work presents an approach based on Symptom Ontologies to facilitate problems identification and solution prediction. The proposal is to use symptom ontology on clickstream analysis, and provide resources for a site to have autonomic actions. To illustrate this idea, two differents symptoms from the clickstream analyses are described. Rules are used to recognize these symptoms, as well their possible effects are analyzed. These effects can correspond to either actions or recommendations that maycorrect a problem.