I've got some data that I need to be normalized. I've tried several different ways I've found online, none of them work the way I'd like. I've tried (x - mean) / standard_deviation, where x is my value. I've also tried (x - min) / (min - max) as seen here. Those two seem to be the 2 standard way of normalizing data that I've seen.

What I'd like is to normalize the data between 1 and 0, cut it off at 3 decimal places, and still have a data make sense. Right now the numbers at the top are so large that they throw everything else off. The first 3 numbers are .8 somethings, but this it quickly drops off to .2 somethings and .1 somethings.

Any suggestions?

  • 1
    $\begingroup$ Would it make sense to look at the Log of the data? $\endgroup$ – Ron Gordon Apr 23 '13 at 13:06
  • $\begingroup$ That helps, leaving the data looking like: pastebin.com/SRtDHpvm $\endgroup$ – sivSivSG Apr 23 '13 at 13:20
  • $\begingroup$ How would you suggest to then make it fall into nice .9** .8** .7** .6** ... buckets? $\endgroup$ – sivSivSG Apr 23 '13 at 13:21
  • $\begingroup$ I'd be more interested in the trend of the data. It looks linear to me - is it? That might help guide you. $\endgroup$ – Ron Gordon Apr 23 '13 at 13:22
  • $\begingroup$ Nah, it's not. There's still this long tail of small values, with a head of larger values above 10. $\endgroup$ – sivSivSG Apr 23 '13 at 13:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.