SSmiRNA

Semi-supervised machine learning integrated pipeline for miRNA prediction. Accompanies "A Semi-Supervised Machine Learning Framework for MicroRNA Prediction" manuscript by Mohsen Sheikh Hassani & James R. Green The "pipeline.py" file demonstrates the general framework of the multi-view co-training process. During each learning iteration, the top two positive and negative instances predicted by one view are added to the initial training set of the other view. The "pipeline.py" file demonstrates the general framework of the multi-view co-training process. During each learning iteration, the top two positive and negative instances predicted by one view are added to the initial training set of the other view. The final train and test sets from the "pipeline.py" file will be the input to the "active_learning.py" file. The "active_learning.py" code is the general method for active learning, but since this is a combination of the two methods, the first part of the code should be commented out, as indicated in the file. "Featureset.py", written by Rob Peace, extracts the sequence and expression-based features. "Smote.py" is used for class imbalance correction