Re:Machine Learning Crash Course: Part 1
by nkozyra on Thursday December 29, 2016 @10:37PM
(#13282324)
In practice it's a mix of mathematical fields; depending on approach, you can have traditional Bayesian probability, regression, estimation, euclidean geometry, classic logic or some combination of the above.
The math is important for understanding and tweaking. But ML is not just fancy statistics (or math in general). It's _data_. An understanding of the math means understanding what data pairs best with what approach. It also means understanding error analysis. It means cross-validation. These are far more vulnerable areas to a beginner in ML.
But the ultimate point is ML is not magic. It's a framework - trenched in mathematics - that provides the building blocks for simulating understanding. You don't need to know the math to use it, but you need it to use it well.
The math is important for understanding and tweaking. But ML is not just fancy statistics (or math in general). It's _data_. An understanding of the math means understanding what data pairs best with what approach. It also means understanding error analysis. It means cross-validation. These are far more vulnerable areas to a beginner in ML.
But the ultimate point is ML is not magic. It's a framework - trenched in mathematics - that provides the building blocks for simulating understanding. You don't need to know the math to use it, but you need it to use it well.