....

January 13, 2021

Why are trees deemed fit for boosting?

Following are the reasons that make trees ideal for Boosting:

  1. Since Decision Trees uses greedy approach, they’re computationally scalable
  2. Robust to outliers
  3. Unaffected by missing values
  4. Feature scaling is not needed
  5. Interpretable for small Boosting models
  6. Work with both categorical and continuous variables

by : Monis Khan

Quick Summary:

Following are the reasons that make trees ideal for Boosting: Since Decision Trees uses greedy approach, they’re computationally scalable Robust to outliers Unaffected by missing values Feature scaling is not needed Interpretable for small Boosting models Work with both categorical and continuous variables