On Friday June 16 Dr. Ravindran gave a talk on gradient boosting. The key points to remember are:

  • Boosting is a technique to create strong classifiers out of weaker ones.
  • The first boosting algorithm, designed in the 80’s?, is Adaboost. It consists on creating some weights for each classifier that contain enough info about their perfomarnce.
  • Great out of the bag classifiers.
  • You can find the slides from the talk here.

Pics from the talk:

⤧  Previous post Gradient Boosting ⤧  Next post SQL for Moving Average - (In-database Analytics)