Extract Rare Itemsets from Large Database Using Automatic Support and Vertical Data Presentation - Rasver
Abstract
The frequent itemset mining is an evolving and surging technology in data mining used to unearth potentially useful hidden patterns present in the database using the user defined support threshold values. When RIM – rare itemsets are extracted from the database using the user defined support threshold, the problem arises due to the following factors. If the minimum support threshold value is set to a high value, the rare itemsets will be ignored and if the minimum threshold value is set to a low value, some uninteresting itemsets will be found. To evade this problem, this research paper introduces automatic support threshold value generation based on the database using an algorithm named “RASVER” and finds the rare itemsets and interesting itemsets without any loss and the time consumption as well as the memory consumption is very low when compared with the existing algorithms.