Algorithmic modeling vs data modeling

In this whitepaper from 2001, Dr. Leo Breiman, the creator of the random forest algorithm, talks about the benefits of a machine learning ("algorithmic modeling") versus a more traditional statistic approach ("data modeling").

The content:

  • His road map
  • Projects in consulting
  • Return to the university
  • The use of data models
  • The limitations of data models
  • Algorithmic modeling
  • Rashomon and the multiplicity of good models
  • Occam and simplicity vs. accuracy

Even considering that the epaper is from 2001, it is still a very interesting read and highly recommended.

Download here.

You will also find PDF's via Google Search, should you not want to give out your email adress.

On Slideshare you will also find a presentation that builds on the pdf.