Dealing with incomplete data based on minimum description length principle
Yongli Li · Journal of Lanzhou University · 2006
The incomplete data is one of the main reason that cause the information system to be uncertain, it causes difficulty in data mining and knowledge discovery. This paper proposed a method for dealing with incomplete data based on minimum description length principle, a case study has been performed, proving its validity, and the rule induction results by Rose tool proved that this method is better than rough sets discernable matrix based method and conditioned mean completer based method in rule concentration and support degree.