Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Meauring deviation of something using some party randomized variables

  1. Jul 4, 2012 #1

    i want to take an x-y data set and find the deviation from another dataset.

    the master dataset is two identical y=1/x curves of different amplitude; also one only registers at periodic intervals and the second is multiplied by a noise signal.

    so part of the problem is going to be that the actual master dataset is inexactly defined as the noise signal itself [i used white noise between 1 and 0] is noise - i.e. i need to be able to say when something is just noise - in which case the match is prefect...

    make sense? can anyone start me off with this [i still need to find the perfect master dataset actually...]. thanks :) !!
  2. jcsd
  3. Jul 4, 2012 #2
    You would probably get better answers if you posted this in the statistics section. You can ask one of the mods to move it for you.
  4. Jul 5, 2012 #3
    trying... :)
  5. Jul 5, 2012 #4

    Stephen Tashi

    User Avatar
    Science Advisor

    You should start by getting your objective clear. "Find the deviation" doesn't define a precise goal. If you want to use statiistical procedures, these fall into two broad categories. There is Hypothesis Testing, which usually considers a yes-or-no type question such as "Are these two data sets measurements of the same underlying phenomenon?" or "Are the mean values of these two distributions identical?". There is Estimation, which considers questions like "What is the mean value of the distribution?" or "What is the numerical difference in the mean values of these two distributions?".
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook