Find a power-series approximation for
and state where it is valid.
Split the logarithm of the quotient first. Rather than expand the whole quotient, use : This is the move that makes the whole problem easy, because both pieces have standard Maclaurin series.
Write down the two standard series. For , and replacing by ,
Subtract and watch the even powers cancel. Term by term, the contributions are and , which cancel on subtraction; the contributions are and , which reinforce. In general only odd powers survive and each is doubled: Only odd powers is exactly what the function's symmetry demands: replacing by inverts the argument and flips the sign of the logarithm, so the function is odd.
Record the useful truncations. Cutting the series off gives the practical approximations The first is the linear (tangent-line) approximation at .
State the interval of validity and test it. Both parent series need , and the quotient itself is only positive there, so the expansion holds for . Numerical check at : the exact value is , while - agreement to five decimal places, with the next term accounting for the small gap.
Note why this form is preferred for computing logarithms. Any positive can be written as with , and that is small even when is far from . Because the series has only odd terms and converges twice as fast as the plain series, this identity is the classical way to get logarithms to high precision by hand.
Need to solve a different problem like this? Open the solver →