Decision Trees for Load Analysis of Computer Networks

The paper elaborates the development of a decision system by means of the decision trees technique. Specifically, classification rules are exported for the loaded state of exterior tine of the computer network of laboratories of Informatics and Computer Technology Department of the Technological Educations lInstitute (TEI) of Lamia/Greece. The decision trees method uses a learning set, consisting of four important attributes of the operation line, the volume of requested data and data that is stored, as well as the average utilization time of the network line. From the decision tree is derived that the time logged on the network and the volume of transferred data, have a great effect on the exterior line loading.