Mining for Association Rules by 4ft-Miner.
Jan Rauch, Milan Šimůnek · 2001
Introduction Procedure 4ft-Miner mines for association rules ' where ' and are derived Boolean attributes. Symbol is 4ft quanti er. Intuitive meaning of ' is that ' and are in relation given by the symbol . Association rule ' is veri ed using four-fold contingency table of ' and . There are association rules corresponding to statistical hypotheses tests see section 2. 4ft-Miner mines also for conditional association rules of the form ' =, see section 3. An example of 4ft-Miner application is in section 4. Main features of 4ftMiner are summarised in section 5. Procedure 4ft-Miner does not use A-priori algorithm. Implementation of 4ft-Miner is based on suitable strings of bits see section 6. Various theoretical results are related to 4ft-Miner see section 7. Let us remark that 4ft-Miner is a GUHA procedure in the sense of [2]. 2 Association Rules Association rule is an expression ' where ' and are derived Boolean attributes. Boolean attributes ' and correspon