2.2 KiB
2.2 KiB
logaddexp
Evaluates the natural logarithm of
exp(x) + exp(y).
Log-domain computations are commonly used to increase accuracy and avoid underflow and overflow when very small or very large numbers are represented directly as limited-precision, floating-point numbers. For example, in statistics, evaluating logaddexp() is useful when probabilities are so small as to exceed the normal range of floating-point numbers.
Usage
var logaddexp = require( '@stdlib/math/base/special/logaddexp' );
logaddexp( x, y )
Evaluates the natural logarithm of exp(x) + exp(y).
var v = logaddexp( 90.0, 90.0 );
// returns ~90.6931
v = logaddexp( -20.0, 90.0 );
// returns 90.0
v = logaddexp( 0.0, -100.0 );
// returns ~3.7201e-44
v = logaddexp( NaN, 1.0 );
// returns NaN
Examples
var incrspace = require( '@stdlib/array/incrspace' );
var logaddexp = require( '@stdlib/math/base/special/logaddexp' );
var x;
var v;
var i;
var j;
x = incrspace( -100.0, 100.0, 1.0 );
for ( i = 0; i < x.length; i++ ) {
for ( j = i; j < x.length; j++ ) {
v = logaddexp( x[ i ], x[ j ] );
console.log( 'x: %d, y: %d, f(x, y): %d', x[ i ], x[ j ], v );
}
}