Efficient Closed Frequent Itemset Mining: Enhancing Association Rule Discovery with Vertical Format Algorithms

Frequent Itemset (FIS) mining has traditionally involved redundant data, but the shift to mining ClosedFIS(CFIS) has mitigated this issue. Despite this improvement, performance challenges such as processing time and storage space persist. New algorithms for CFIShave been introduced, and research in Association Rule Mining (ARM) remains active, mainly focusing on efficient itemset discovery and rule optimization. This has led to well-established algorithms, including Apriori, Apriori-Tid, H-Mine algorithms but this research concentrated on vertical format algorithms because of Apriori property exploitation for generating candidate itemsets. This paper shows that mining all FIS initially isn't necessary; instead, focusing on CFIS, which are significantly fewer, is sufficient. It also eliminates the need to mine every possible rule, as any rule between itemsets can be represented by a rule among closed itemsets. The equivalence class transformation algorithm focuses on identifying frequent items within transaction groups, operating exclusively on a vertical database layout to minimize redundant rules. Additionally, the Improved Closed Association Rule Mining (ICHARM) method is introduced as an efficient method for extracting all closed frequent itemsets. Experimental results demonstrate that this approach can significantly enhance sales strategy development.