Fig. 1

Imputation Process. This figure illustrates how the imputation and the feature ranking process are connected. At first, the imputation models are built using features, which show a minimum correlation (here 7% was used) to the feature to be imputed. After that, the obtained 5 different data sets are used for the feature selection process. Knowing the selected features, the imputation is re-done. This by using as predictors additional to the correlated features also the selected ones