Efficient Mining of Association Rules From Merged Itemsets

• Main Authors: Shi-Qin Weng, 翁世親
• Other Authors: Dong-Lin Yang
• Format: Others
• Language: en_US
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/r3e978

碩士 === 逢甲大學 === 資訊工程所 === 91 === Mining for association rules between items in a large database of sales transactions has been widely investigated recently. In this thesis we use three kinds of data preprocess methods to merge itemsets in a database and present an algorithm to mine association rules from merged itemsets. One of data preprocesses method is using sort to group the similar transactions. The other two are used a simplified dynamic programming algorithm to merge transactions. Compared to the other algorithms for mining association rules, our methods reduce a lot of I/O overhead. The results of our experiment show that it only costs one pass of scanning database. Therefore, our methods are especially suitable for very large size databases. They are also ideal for mining association rules with updated databases.