Loss of significance is an undesirable effect in calculations using floating-point arithmetic. It occurs when an operation on two numbers increases relative error substantially more than it increases absolute error, for example in subtracting two nearly equal numbers (known as catastrophic cancellation). The effect is that the number of accurate (significant) digits in the result is reduced unacceptably. Ways to avoid this effect are studied in numerical analysis.
Read more about Loss Of Significance: Demonstration of The Problem, Workarounds, Loss of Significant Bits, Instability of The Quadratic Equation, A Better Algorithm
Famous quotes containing the words loss of, loss and/or significance:
“The loss of enemies does not compensate for the loss of friends.”
—Abraham Lincoln (1809–1865)
“The loss of enemies does not compensate for the loss of friends.”
—Abraham Lincoln (1809–1865)
“It is necessary not to be Christian to appreciate the beauty and significance of the life of Christ.”
—Henry David Thoreau (1817–1862)