The Search for Twitter Spam Bots

The Search for Twitter Spam Bots

This project includes finding spammer and bot accounts in Twitter using machine learning techniques. The dataset is provided at the Kaggle website and I have compared the performance of the following algorithms:

  • ID3 Decision Tree
  • CART: Classification and Regression Trees
  • Gaussian Naive Bayes
  • SVM
  • Logistic Regression
  • SVM + CART
  • Bagged Forest CART

The results show that Bagged Forest Cart algorithm and CART has the best accuracy on finding spam account.

Accuracy of different algorithms.

Avatar
Amir Yazdani
PhD Candidate in Robotics