Build Classifier

This wiki provides details about using buildClassifier.py script.

Required Dependencies

Python 2.x and following python modules:

• pandas
• Numpy
• Scipy
• Matplotlib
• scikit-learn - This python package provides variety of machine learning algorithms. Downlaod - http://scikit-learn.org/stable/install.html#

Script Parameters

n_trees- The number of trees in the forest (Default - 100)
n_features- The number of features to consider when looking for the best split (Default - 5)
n_classes- Binary class classification (Default - 2, Do not change)
n_folds- Number of k-fold cross-validation (Default - 5)
n_jobs - The number of jobs to run in parallel (Default - 10)
Best_k_features - Number of features to select using Information gain (Default - 250)
InputFile- Feature file generated from Generate_Features.py (e.g. 'Features.csv')

Running the script

python buildClassifier.py

Output File

• Results Folder
  1. Performance_Report.txt - Performance measures such as accuracy, sensitivity, specificity, precision, confusion matrix, and information gain based ranked Best_k_features
  2. ROC_Detailed.png