The goal of this task is to train a classifier to detect political bias in news articles.

The data folder contains three files

• bias_articles_train.json: Json containing array of dicts with
  • id: unique integer id for each article
  • title: Title of the article
  • body: Body of the article
  • bias: Class label it can be "Left", "Center" and "Right"
• bias_articles_dev.json: Same as bias_articles_train, but used for validation of the models.
• bias_articles_test.json: Same as bias_articles_train, but does not contain bias labels (to be predicted.)

• In bias_classifier.py Implement fit, test and predict functions.
• You may use any classifier that you like, any features extracted from the title and the body.
• Use bias_articles_train.json to train the model, bias_articles_dev.json to fine-tune the parameters.
• Predict the labels for articles in bias_articles_test.json using the model trained above.
• Document the code and follow Google python coding conventions
• Document the results of your model (train performance vs test performance) in Results.md markdown file.