期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2015
卷号：3
期号：11
DOI：10.15680/IJIRCCE.2015.0311075
出版社：S&S Publications
摘要：The data mining techniques are employed in library reader management system for providingpersonalized services to readers. The data used in mining for this purpose mostly will be of categorical in nature. Thestudy was undertaken to discover an efficient categorical clustering algorithm which could replace the k-meansnumerical clustering method in the current mining system.QROCK or Quick RObust Clustering using linKs [2] algorithm was found to be most suitable for categorical clusteringbecause of its run time efficiency and simplicity. It is an agglomerative hierarchical clustering method which computesclusters by determining number of connected components in a graph and could replace many overheads in k-meansclustering[8]. Categorical data itself can be given as the input which could eliminate categorical-numerical conversionfor its kind. Algorithm can produce desired number of clusters unlike k-means and input overheads in the form of kvaluegiven by the user prior to its execution have also been eliminated. This study identified that QROCK algorithmcan be used for categorical clustering in library reader management system efficiently
