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:
“Eventually we will learn that the loss of indivisible love is another of our necessary losses, that loving extends beyond the mother-child pair, that most of the love we receive in this world is love we will have to share—and that sharing begins at home, with our sibling rivals.”
—Judith Viorst (20th century)
“Mothers risk alienating their mates if they expect them to hold or care for the baby exactly as they do. Fathers who are constantly criticized or corrected may lose interest in handling the baby, and this is a loss for everyone. The cycle is a dangerous one. Now the same mother feels bitter because she is no longer getting any help at home.”
—Cathy Rindner Tempelsman (20th century)
“Politics is not an end, but a means. It is not a product, but a process. It is the art of government. Like other values it has its counterfeits. So much emphasis has been placed upon the false that the significance of the true has been obscured and politics has come to convey the meaning of crafty and cunning selfishness, instead of candid and sincere service.”
—Calvin Coolidge (1872–1933)