Simple articles topics classifier that utilizes SVM machine learning algorithm.

Based on scikit-learn and Python 3. Only binary classification is supported in this version.

Installation and running

1. You need some articles data to train and classify to be stored at data directory. Classifier assumes that you have 3 subdirectories:

   data/classifier_train_data/<class_name>/<article>.txt   # ML algorithm training data
   data/classifier_test_data/<class_name>/<article>.txt    # Classified test data
   data/classifier_x_val_data/<class_name>/<article>.txt   # All available articles data for classifier cross-validation
   

2. Install NumPy and SciPy. It's highly recommended by NumPy authors to install binaries system-wide. So, we won't use virtualenv.

   sudo apt-get install python3-numpy python3-scipy
   

3. Install required Python packages

   pip3 install -r requirements.txt
   

4. Running

   python3 text_classifier.py