Topological frequency utility mining model

Abstract

Most of the traditional association rules mining only considers frequency of itemsets (web pages set denoted as wps-set in a given web site), which may bring less amount of profit. The main focus of Utility mining is to consider only itemsets (wps-set) with high utilities, but number of rich enough customers is restricted. To overcome the weakness of the frequency and utility mining models, General utility mining model was proposed [7]. Even though General utility mining focuses both on frequency and utility, the location/position of item (web page) is not taken into account. To overcome weakness General Utility Mining model, we propose a novel model called Topological Frequency Utility Mining Model which takes along with frequency and utility the position of web page (item), at what distance/level it is from home page (entrance of the shop), number of forward & backward links (number of items stacked in a particular partition in the shop) in the web site. In this work a metric is proposed, which aim to quantify the importance of a web page based on frequency, utility and topology of the web site. Subsequently certain guidelines are presented which may help web site designers to restructure the web site taking into account the optimization of metric. © 2012 Springer India Pvt. Ltd.