Missing Data Testing is a collection of R scripts for testing imputation accuracy and computation time using a variety of imputation techniques as described in JBHI-00775-2020
Download and Run. Requires the ADNIMERGE package (adni.loni.usc.edu) to be installed in R if using the same data as in the paper.
Create a new R project and run the scripts. Steps 0-4 are for imputation accuracy and time testing ( results shown in Fig. 4 in paper)
#Step 5 onwards is for the imputation + classification LOOCV accuracy and time testing (results shown in Fig. 5 in paper)
#MATLAB code for Figure 3 and 4 is included.
#an slight modification of the missForest method in R which allows for the specification of a test set to be imputed separately from the training set is also included