- Type: System..::..DoubleOn entry: the argument of the function.
Abramowitz M and Stegun I A (1972) Handbook of Mathematical Functions (3rd Edition) Dover Publications
Because of its close relationship with , the accuracy of this method is very similar to that in s15ad. If and are the relative errors in result and argument, respectively, they are in principle related by
so that the relative error in the argument, , is amplified by a factor, , in the result.
For small and for positive this factor is always less than one and accuracy is mainly limited by machine precision.
For large negative the factor behaves like and hence to a certain extent relative accuracy is unavoidably lost.
However the absolute error in the result, , is given by
so absolute accuracy can be guaranteed for all .
This example reads values of the argument from a file, evaluates the function at each value of and prints the results.