Null Value Estimation Method Based on Information Granularity for Incomplete Information System

In actual life, there are lots of incomplete information systems. One of the way to deal with incomplete information system is to complete the null value using estimation methods. Traditional estimation methods is mainly based on the appearing frequency of other values with same attribute and the value estimated by these methods maybe not get the best classification result, thus it leads to a lower support degree and confidence degree, so increases the uncertainty of information system. This paper proposed a null value estimation algorithm GRCC based on information granularity. The uncertainty degree of the information system is measured by information granularity, choose the attribute whose information granularity is the max in left attributes and complete the null value of this attribute. Repeat this program until the whole incomplete information system is completed. Test results shows that the uncertainty degree of incomplete information system can be decreased effectively.