Serum_ML

The script final_predictor_kbest.py performs a feature selection by Recursive Feature Elimination with Cross-Validation, and utilizes the selected features for training the serum classifier. Run the script after cloning without any parameters. An output folder will be created with the results. The classifier is trained on 54 samples and its performance is tested on 33 samples. The script utilizes the protein intensity matrix LFQ_intensity.txt and the experimental design file expdesign.txt from the data folder to train and test the classifier. The output folder contains the roc_curve, confusion_matrix and rfecv plot along with the prediction result and several other tables.