0
$\begingroup$

Maybe there is already a question similar to mine but there are so many involving the term accuracy and at least none, except of How to evaluate instrumentation accuracy? , didn't seem "very" similar.

A simple question: The task is to measure a device and f(x) has to be determined under a given accuracy. For example, there is a voltage measurement device and you have to check if the output voltage is better than 0.1 V precise. How to proceed that? Is it possible one with the help of the data points? I assume it isn't that trivial as probably several errors have to be taken into account? EoV maybe, too?

$\endgroup$
5
  • 1
    $\begingroup$ This is very vague. Are you asking about a confidence interval? $\endgroup$ Commented Jul 15, 2019 at 7:04
  • $\begingroup$ Maybe, I forgot to ask if a confidence interval would represent an estimation accurancy? Means, I would need to fulfill a given confidence interval instead of performing some methods and gather which confidence intervals they do have. Is that even possible? Or do I have to trial and error and hope that the confidence interval is as required? $\endgroup$
    – Ben
    Commented Jul 16, 2019 at 6:18
  • $\begingroup$ A confidence interval would tell you the range of possible values of your estimate. So for ex. if your estimated accuracy would be 0.5 and your 95 % CI would be [0.4,0.6], that would give you a sort of range of likely possible values. $\endgroup$ Commented Jul 16, 2019 at 7:36
  • $\begingroup$ thanks, but I won't know this before I did everything already, or? $\endgroup$
    – Ben
    Commented Jul 16, 2019 at 8:07
  • $\begingroup$ You estimate the CI after the fact, just like any other estimate. Maybe you are looking for some sort of power analysis? $\endgroup$ Commented Jul 16, 2019 at 12:10

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.