Rehabilitating of incomplete data sets based on Bayesian networks

The problem of repairing incomplete data sets for missing data in the Bayesian network is investigated in this paper. The paper provides a method, which is the neighboring distance regression, to repair incomplete data sets. This method first repair the incomplete data sets to complete one, then compute with them, and treat the results as the approximation of the incomplete ones. Experimental results show that this method has exact correct rates and can gain good effects with large database. This method is perfect to deal with missing data.