A novel approach to attribute reduction and rule acquisition of formal decision context

Rule acquisition and attribute reduction are important research topics in formal concept analysis. Many existing rule-based attribute reduction algorithms are designed to computing all reductions by using discernibility functions and therefore these algorithms are NP-hard. To improve the applicability of rule-based attribute reduction algorithms, firstly, we propose a method to simplify the discernibility matrix such that fewer concepts need to be distinguished. Then a heuristics approach is presented to compute one reduction by using the ordered attributes method. In addition, a novel rule acquisition algorithm for OW-decision rules is presented. Some comparative analyses of the rule acquisition algorithm with the existing algorithms are examined which shows that the algorithms presented in this study behave well. And finally, we select some datasets from UCI datasets for taking experiments and illustrate the effectiveness and efficiency of our proposed reduction algorithms.