>>14407823Data fitting and estimation for the most part. I don't really like the term machine learning. I prefer statistical learning.
Think of linear regression (fitting a line through data) as the most basic problem in machine learning. All other algorithms are advanced forms of that problem. Some handle high dimensions well (i.e. finding f(x,y,z....)) Some handle large amounts of data well some are focused classification (i.e. f(x) = 1 or 0) others are continuous focused. Some are better at time series data, some are focused on knowing the underlying model first, etc.
You get the point. It's a huge field because prediction and interpolation are big money making ventures and the problems are pretty vast because data can take on so many forms. There really isn't a one size fits all solution either and many of the algorithms can fuck up.
