How to check the accuracy of estimation

5 views (last 30 days)
Lorenne
Lorenne on 7 May 2018
Edited: Stephen23 on 7 May 2018
Using the approximated volume i have 0.51666667 and the true volume is 0.52359878 , the error is 0. 00 693211 so it will be accurate by 2 decimal places. How do i only print the values if only the decimal place has increased by 1? I used floor(-log10(error)) but it counted incorrectly as the error gets smaller. For points of 168, the zeros in front should be 5, but it calculated to be 6 . May i know what is wrong?
  1 Comment
Stephen23
Stephen23 on 7 May 2018
Edited: Stephen23 on 7 May 2018
"For points of 168, the zeros in front should be 5, but it calculated to be 6 . May i know what is wrong?"
Look more closely at the error value and you will find that it actually has 6 leading zeros. Try fprintf to thirty significant figures, or download and use James Tursa's excellent num2strexact.
Welcome to the world of floating point numbers! Note that how floating point numbers are displayed is not the same thing as what value is actually stored in memory.

Sign in to comment.

Answers (1)

Walter Roberson
Walter Roberson on 7 May 2018
Observe:
>> fprintf('%.8f\n', 0.000000999)
0.00000100
Actual number of leading 0: 6. The 5 visible are due to rounding for presentation.

Categories

Find more on Tables in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!