Maximal association analysis using logical formulas over soft sets

Abstract

Discovering interesting and useful association rules from the collected data is an issue of great importance in pattern mining. Although a myriad of association rules can be extracted with traditional rule mining techniques, some of the obtained rules might be redundant or even meaningless in many cases. To overcome this difficulty, logical formulas over soft sets are applied to maximal association mining in this study. With the help of logical formulas over soft sets, all critical concepts for mining both regular and maximal association rules are incorporated into a common framework, and uniform mathematical characterizations of these concepts are provided accordingly. Three algorithms are also designed to develop a new approach to maximal association rule mining using logical formulas over soft sets. Moreover, we present two examples to show theoretical value of the obtained results and practical applicability of the proposed approach. The first example relies on a clinical diagnosis data set and illustrates the advantages of applying logical formula over soft sets to rule extraction. In the second example, we conduct a case study based on a data set regarding Nobel Laureates to show that the proposed approach is helpful for discovering interesting facts in real-world scenarios.