Modified APRIORI Technique For Association Rule Mining And Its Effectiveness In Prediction

Abstract

Data mining is a common methodology used for mining web pages. It is performed to basically collect all the historical data present in server log files. Server log repository is the place where all the data concerning the website, is stored as weblogs. In order to analyze the collected data through data mining several measures are to be undertaken like preprocessing, path completion, clustering, pattern discovery and analysis. The steps include cleaning, user identification, session identification and transaction identification. Association rule learning can be used to discover relations between variables in large datasets. It can identify strong rules using some measures of interests. Frequent item sets are generated by framing association rules that will lead to buying or reaching a particular webpage. Apriori algorithm is used for identifying frequent item sets from the data set. Modified Apriori technique is used here to show its supremacy over conventional Apriori. In this article, a study about modified Apriori technique that can be used to effectively predict next web pages is discussed.