Enhanced K-Means Clustering Algorithm for Evolving User Groups

To gain information about user interests in Web pages is needed to advance in Web security. An approach to pick up that information includes understanding the client's perusing conduct, examining the Web log records with the procedures of preprocessing and client clustering. Time spent on Web pages and the types of operations show the degree of a Web user's intention. The data set comprises of Web log files obtained by collecting the user logs during a six month period. A new enhanced K-means clustering algorithm proposed in this paper for grouping user based on their preferred Web content and their temporal constraints. The enhanced K-mean clustering calculates initial centroids instead of random choice and uses time intervals to heighten the security and performance. Utilizing this methodology, client access designs with comparable looking practices are assembled into a particular class amid a particular time interval. Also secured communication among the various users groups will be achieved through hill cipher technique.