• use simple models as base models
  • speeds up training and inference time
  • without necessarily impacting the performance of the ensemble
• can learn complex functions
• Noise and outliers should be removed from data
• Prefer simple models over complex models

Built with DataHub Cloud