Rough-set-based approaches to data containing incomplete information: possibility-based cases
Michinori Nakata, Hiroshi Sakai · 2005
Methods based on rough sets to data containing incomplete information are examined under a possibilty-based interpretation for whether a correctness criterion is satisfied or not. The correctness criterion is to give the same results as methods by possible tables. The methods proposed so far do not give the same results as methods by possible tables. Therefore, we show a new formula not using implication operators in methods by valued tolerance relations. The formula bears the results that agree with ones from using possible tables. Thus, by using the formula the methods by valued tolerance relations satisfy the correctness criterion.