J48 Tree Classification and Simple K-Means Clusterization on the Behavior of Academic Community in Accessing Internet for Bandwidth Management Plan

Abstract

<p>Abstract The purpose of the research was to review the<br />improvements in internet usage behavior by the academic<br />community at a university which consisted of favorite website<br />variables, the amount of internet bandwidth capacity used and<br />the time spent accessing the website. The research was<br />conducted in a population of 200 lecturers and 5267 students.<br />Internet bandwidth mapping is carried out using two methods of<br />the J48 tree classification algorithm and simple K-Means<br />clustering contained in the WEKA software. The results of this<br />study are for the purpose of mapping internet bandwidth,<br />determining internet service provider load balancing,<br />classification and clustering of favorite websites based on time<br />duration which can affect the use of existing internet bandwidth<br />capacity systems.</p>