What Is A Floating Point In R programming

How floating points influence machine learning

A floating point can seem like a very nuanced academic topic for creating a data model. However, it is a concept that has become important as data calculations involving decision-influencing machine learning have been increasing over the years. Developing a small appreciation for it can help you maintain data model quality.

According to Wikipedia, a floating point is a type of arithmetic meant to approximate a real number. It is designed to provide a level of computing accuracy that can be managed within a given amount of computer memory.

A number as a floating point consists of a significand, multiplied by a base and exponent.

So let’s take the following number, 1.2345, from the perspective of a floating point.

The number 1.2345 can be expressed as 12345 times 10 to the negative 4th power. The significand is the integer, 12345. 10 is the base integer, while -4 is the exponent. The definition floating point refers to the “float” of the decimal in its assignment within a real number.

The significand, exponent, and base are stored in computer memory as a sign, exponent, and fraction. Each bit is meant to cover a portion of these facets. A sign is usually one bit, exponent 8 or 16 bits, and a fraction makes up the rest. It is these bits that denote the capacity for accuracy.