Goals

• Use SFrames to do some feature engineering.
• Modify the decision trees to incorporate weights.
• Implement Adaboost ensembling.
• Use your implementation of Adaboost to train a boosted decision stump ensemble.
• Evaluate the effect of boosting (adding more decision stumps) on performance of the model.
• Explore the robustness of Adaboost to overfitting.

Packages used

• graphlab
• matplotlib

Used data set

lending-club-data.gl

Algorithms used :

• decision trees.
• Adaboost ensembling.
• boosted decision stump ensemble.