4 Model Fitting

In this project, seven different machine learning models were fit in an attempt to classify fake news. The models used were:

  • Naive Bayes
  • Basic Logistic Regression
  • Lasso Regression
  • Support Vector Machine
  • Random Forest
  • Multilayer Perceptron
  • Recurrent Neural Network

The first five models were fit in order to get a baseline sense of how non-deep learning algorithms performed when classifying fake news. Then, the two deep learning models were created in an effort to improve on the previous models. Since this process requires quite a bit of code as well as additional explanation for each model, it has been moved to the Appendix in Section 7. More specifically, the entire process of model fitting (with code) can be found in Sections 7.2 and 7.3 for the baseline models and deep learning models respectively.