• Maximum Entropy Models handle overlapping features well
• unlike Naive Bayes Model, there is no double counting
  
• Maximum Entropy Models handle overlapping features well, but do not automatically model feature interactions. They do not model interaction terms, you have to add them