Application of MapReduce to Ranking SVM for Large-Scale Datasets