3.1 KiB
3.1 KiB
expm1rel
Compute the relative error exponential.
The relative error exponential is defined as
Usage
var expm1rel = require( '@stdlib/math/base/special/expm1rel' );
expm1rel( x )
Computes the relative error exponential.
var v = expm1rel( 0.0 );
// returns 1.0
v = expm1rel( 1.0 );
// returns ~1.718
v = expm1rel( -1.0 );
// returns ~0.632
v = expm1rel( NaN );
// returns NaN
Notes
- When
x
is near zero,exp(x)-1
can suffer catastrophic cancellation (i.e., a significant loss in precision).expm1rel
avoids such a loss in precision.
Examples
var randu = require( '@stdlib/random/base/randu' );
var expm1rel = require( '@stdlib/math/base/special/expm1rel' );
var x;
var y;
var a;
var i;
for ( i = 0; i < 100; i++ ) {
x = (randu()*100.0) - 50.0;
a = x.toFixed( 3 );
y = expm1rel( x );
console.log( '(e^%d - 1)/%d = %d', a, a, y );
}