期刊名称:International Journal of Computer Science and Network Security
印刷版ISSN:1738-7906
出版年度:2008
卷号:8
期号:8
页码:195-200
出版社:International Journal of Computer Science and Network Security
摘要:Most of the association rules mining algorithms to discover frequent itemsets do not consider the components of transactions like its quantity or weight or its total profit. In a large database it is possible that even if the itemset appears in a very few transactions, it may be purchased in a large quantity for every transaction in which it is present and may lead to very high profit. Therefore the weight is the most important component and without which it may lead to lose of information. Our novel method discovers all frequent itemsets based on its weights in a single scan of the database. In order to achieve this, we first construct a weighted tree containing weights and a set of transaction id’s corresponding to every attribute in the database. Then by scanning the above tree we can discover all frequent itemset. This method is also found to be efficient than FP-tree, which require two scans of the database to discover all frequent itemsets based on user defined minimum support. Further, using Weighted tree constructed for the above method is used for discovering best order or sequence of attributes. If database is read in this best sequence , we can construct an abstraction of the entire database in the memory which occupies less space in the memory when compared to sequential reading.
关键词:Association rule, Data mining, frequent itemsets, Support, Weighted Support, Weighted tree