Testwiki:Reference desk/Archives/Mathematics/2016 August 18

{| width = "100%"

|- ! colspan="3" align="center" | Mathematics desk |- ! width="20%" align="left" | < August 17 ! width="25%" align="center"|<< Jul | August | Sep >> ! width="20%" align="right" |Current desk > |}

Welcome to the Wikipedia Mathematics Reference Desk Archives
The page you are currently viewing is a transcluded archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

August 18

Normalized Mean Square Error

One professor has given me formula for normalized mean square Error(NMSE),

$NMSE=\left[\frac{MSE(t-a)}{mean(variance(t))}\right]$,

where t is target and a is actual column matrix.I dont understand NMSE formula. plz help.me.what should be NMSE value if target is near to actuals. I get NMSE near to 1 when target is near to actual .He says the reverse. plz provide me text book link if anyone identifies this formula.
Thanks in advance.
plz 112.133.223.2 (talk) —Preceding undated comment added 06:30, 18 August 2016 (UTC)

I have no reference text to suggest, but you could look at our article Mean squared error. Your professor's formula gives the normalized form as the MSE divided by a denominator. If the target values equal the actual values, and if I understand your use of those terms, then the errors are all zero, so the squared errors are all zero, so the mean of the squared errors is zero. Thus your numerator is zero, so the entire expression for NMSE is zero. Loraof (talk) 21:30, 18 August 2016 (UTC)

thanks 112.133.223.2 (talk) —Preceding undated comment added 07:34, 19 August 2016 (UTC)