Efficient Algorithm for Discovering Potential Interesting Patterns with Closed Itemsets
A pattern discovered from a collection of data is usually considered potentially interesting if its information content can assist the user in their decision making process. To that end, we have defined the potential interestingness of a pattern based on whether it provides statistical knowledge that is able to affect one’s belief system. In previous work, we proposed two novel algorithms, Discovering All Potentially Interesting Patterns (DAPIP) and All-Confidence Discovery of Potentially Interesting Patterns (ACDPIP), designed to discover potentially interesting patterns from a collection of data. Results of experimental investigations show that the application of these two algorithms is limited to non-dense datasets. In response, we propose a new algorithm, referred to as ACDPIP-Closed, designed to discover potential interesting patterns from dense datasets. We show empirically that ACDPIP-Closed is able to effectively and efficiently discover potentially interesting patterns from dense datasets. Additional contributions provided by the paper include a definition of a frequent closed itemset based on an all-confidence threshold and a theorem stating that, under the assumption of a particular ordering of items, an itemset is support based closed if and only if it is all-confidence based closed.