Definition
In computer systems the type of numeric data approximately representing real numbers. A floating point number is composed of an exponent and a mantissa specifying the numerical value relative to the exponent. Owing to the limited length of a computer WORD, the floating point representation may cause a loss of precision.