An Efficient and Accurate Discovery of Frequent Patterns Using Improved WARM to Handle Large Web Log Data

An Efficient and Accurate Discovery of Frequent Patterns Using Improved WARM to Handle Large Web Log Data

Sahaaya Arul Mary (Jayaram College of Engineering and Technology, Tiruchirappalli, Tamil Nadu, India) and M. Malarvizhi (J.J. College of Engineering and Technology, Tamil Nadu, India)
DOI: 10.4018/ijitwe.2014040103
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

In the booming era of Internet, web search is inevitable to everyone. In web search, mining frequent pattern is a challenging one, particularly when handling tera byte size databases. Finding solution for these issues have primarily started attracting the key researchers. Due to high the demand in finding the best search methods, it is very important and interesting to predict the user's next request. The number of frequent item sets and the database scanning time should be reduced for fast generating frequent pattern mining. It fulfills user's accurate need in a magic of time and offers a customized navigation. Association Rule mining plays key role in discovering associated web pages and many researchers are using Apriori algorithm with binary representation in this area. But it does not provide best solution for finding navigation order of web pages. To overcome this, weighted Apriori was introduced. But still, it is difficult to produce most favorable results especially in large databases. In the effort of finding best solution, the authors have proposed a novel approach which combines weighted Apriori and dynamic programming. The conducted experiments so far, shows' better tracking of maintaining navigation order and gives the confidence of making the best possible results. The proposed approach enriches the web site effectiveness, raises the knowledge in surfing, ensures prediction accuracies and achieves less complexity in computing with very large databases.
Article Preview

Introduction

Web search engines have been used to the highest degree in online services. Millions of searches performed each second in a Web and their popularity is terrific based on their simplicity of usage. Search information is stored in server log and these log data are huge stream data. The stream data is an ordered sequence of items that arrives in timely order. These data are different from traditional static databases and they are continuous, unbounded, usually come with high speed and have a data distribution that often changes with time. As the number of applications on mining data streams grows rapidly, there is an increasing need to perform association rule mining on stream data in web.

Association Rule is an essential topic in the research of knowledge discovery in Data Mining. The Association Rule Mining received a significant research attention. Association Rule Mining (Soriris, 2006) is one of the strategies that find out association among the pages visited together frequently. Agrawal et al. (1993) first introduced the ARM for Market Basket analysis Problem. Association rule mining has a wide range of applicability such as medical diagnosis, medical research, website navigation analysis, homeland security and so on. For the past few years, researchers proposed a large number of efficient Association Rule Mining algorithms (Renata, 2006). Among these, the Apriori algorithm gains more popularity. The Apriori algorithm (Veeramalai, 2010) has not only influenced the association rule mining community, but its impact is also on other areas of data mining. It is the best-known algorithm to mine association rules. The breadth-first search strategy is used in Apriori to count the support of item sets and also used for a candidate generation function which exploits the downward closure property. Apriori property states that all the subsets of frequent item sets must also be frequent. The researchers, to improve the performance of the Apriori algorithm, did various research works. The variations in Apriori algorithm are Partitioning, Sampling, Tree projection and FP-Growth algorithm. The improved Apriori (Prakash, 2010) evaluated scalability of the algorithm by considering transaction time, number of item sets used in the transaction and memory utilization.

Generalized Association Rule Mining (Mohd, 2008) algorithm reduces the number of rules by pruning and filtering the frequently accessed rules. Distributed Association Rule Mining (Zheng, 2010) algorithm applied in a distributed database and it is a matrix-based algorithm affords less time complexity, space complexity and cost. Fuzzy Class Association Rule Mining (Shingo, 2011) method is based on genetic networks programming and is designed to handle both discrete and continuous attributes. A new technique (Fengjun, 2007) such as Combined Association Rule Mining Algorithm joined with hash tree and genetic algorithm was developed to solve Genes with similar patterns of RNA. A novel method (Victoria, 2011) for mining association rules from semantic instance data repositories is expressed in RDF and OWL. It is used to feed traditional association mining algorithm. Pareto based Multi Objective evolutionary algorithm(Peter,2008) is used for prediction accuracy, comprehensibility and interestingness. It seems to satisfy a different objective of Association Rule Mining problem. Yue Xe et al. (2011) presented reliable representations in association rules for eliminating the redundancy in large web data. They introduced the frequent closed item set instead of frequent item sets. Other Association mining techniques are Mining Frequent Sequences, K-Optimal Pattern Discovery, Quantitative Association Rules, Interval Data Association Rules, Maximal Association Rules, Filtered Associations, Predictive Apriori and Tertus.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 13: 4 Issues (2018): 1 Released, 3 Forthcoming
Volume 12: 4 Issues (2017)
Volume 11: 4 Issues (2016)
Volume 10: 4 Issues (2015)
Volume 9: 4 Issues (2014)
Volume 8: 4 Issues (2013)
Volume 7: 4 Issues (2012)
Volume 6: 4 Issues (2011)
Volume 5: 4 Issues (2010)
Volume 4: 4 Issues (2009)
Volume 3: 4 Issues (2008)
Volume 2: 4 Issues (2007)
Volume 1: 4 Issues (2006)
View Complete Journal Contents Listing