A Novel Hybrid Approach to Estimating Missing Values in Databases Using K-Nearest Neighbors and Neural Networks
Missing values in datasets and databases can be estimated via statistics, ma-chine learning and artificial intelligence methods. This paper uses a novel hybrid neural network and weighted nearest neighbors to estimate missing values and provides good results with high performance. In this work, four different characteristic datasets were used and missing values were estimated. Error ratio, correlation coefficient, prediction accuracy were calculated between actual and estimated values and the results were com-pared with basic neural network-genetic algorithm estimation method.