Fig. 1

Flowchart for the predictive model construction and performance evaluation. First, we randomly divided the HapMap dataset into the training set (85 % of samples from each population) and independent test set (15 % of samples from each population). Then, the training samples were further partitioned into 10 equally sized partitions for 10-fold cross validation. Based on the training dataset, the features were selected, and the predictive model was constructed. Finally, the constructed model was tested on the independent test dataset