Handling missing Data values in a Database Model using Random Forest

Missing values in a databases one of critical problem faced by the researchers in Data analysis and data mining. This work presents a suggested method for handling missing data values in data sets using Random Forest (RF) Technique. The use of RF present new principles to random splitting, it alters the tree growing process by narrowing its focus during split selection. For example, if the database contains numbers of columns usable for prediction, RF would begin randomly of selection number of variables and then chooses the splitter from the list of predictors. Using the suggested method we can get the actual values for the missing records entries and handling the uncertainty and outliers problem.