A Supervised learning algorithm which can classify multi-class problems. Brief Description: A random forest classifcation algorithm which used Kmeans ro perform data splits at each node ( the splits are unsupervised , and donot use labels)

• Data format: H (n*m) input matrix with n points and m features .

• Labels (n*1) - Labels to corresponding data files.

• Creating Train and Test datasets for multiple runs. You can use the "supervised_bagging" function of this purpose. (Let me know if any difficulty)

• TrainDMapIndices - Indices of randomly shuffled training points.

• TrainLabels - Corresponding training labels.

• TestDMapIndices - Indices of randomly shuffled testing points.

• TestLabels - Corresponding testing labels.

• Other parameters in the file can be changed according to the need.

Please let me know before any major changes.