This code uses machine learning in the form of two PU Learning techniques (Two-Step and PU Bagging) and two One-Class Classification techniques (One-Class SVM and Isolation Forest) for predicting functional and non-functional targets in the problem of miRNA target Prediction. Two supervised methods (Random Forest and SVM) are also used for comparison to the results obtained with PU Learning and OCC.
- python=3.7
- pandas
- numpy
- matplotlib
- scikit-learn
- imbalanced-learn
The dataset used can be found at https://drive.google.com/open?id=1SPVYiqNMeOiwFasTUHtiFDx81ji_xCYb
Only the .att file is needed and it should be added to a "datasets" folder inside the main folder.
All files generated on each run will be stored inside a "executions" folder.
The main script file that should be run is the "tarbasePU.py" file.