Smartcore

Machine learning and SmartCore

Machine learning is seen as equal parts buzzword and obscure science, sometimes frowned upon as a new trend, destined to fade away.

In the simplest possible terms, machine learning is a series of algorithms that allow the computer to learn from studying data and statistics.

At its core, then, machine learning is data manipulation within the framework of statistics.

Obviously, we can’t explain everything about how machine learning works in a single lesson, because it involves quite a bit of knowledge of math and statistics.

Instead, we’ll explore SmartCore, a library suited for both machine learning and numerical analysis.

SmartCore has a collection of prebuilt algorithms that can be applied generically in machine learning. It is not a tool for experts, but its algorithms need very little adaptation for most of the tasks for which a generic user might need machine learning.

The basics of SmartCore

SmartCore covers many algorithms and tasks in machine learning. For our purposes, we’ll concentrate on two cases:

• Classification: Given some data, the algorithm predicts a class for the data. For example, we provide some characteristics of a flower, and the algorithm has to tell which flower it is. In another case, we give it the results of a medical test, and it predicts whether ae patient is positive or negative for a disease.

• Regression ...