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:
“Children, dear and loving children, can alone console a woman for the loss of her beauty.”
—Honoré De Balzac (1799–1850)
“Unthinking people will often try to teach you how to do the things which you can do better than you can be taught to do them. If you are sure of all this, you can start to add to your value as a mother by learning the things that can be taught, for the best of our civilization and culture offers much that is of value, if you can take it without loss of what comes to you naturally.”
—D.W. Winnicott (20th century)
“For a parent, it’s hard to recognize the significance of your work when you’re immersed in the mundane details. Few of us, as we run the bath water or spread the peanut butter on the bread, proclaim proudly, “I’m making my contribution to the future of the planet.” But with the exception of global hunger, few jobs in the world of paychecks and promotions compare in significance to the job of parent.”
—Joyce Maynard (20th century)