How to form the input values for the NS correctly. - page 2

New comment
 
Why the first one?
 

If normalised with respect to the first value, the first value will always be zero. Another option is to normalise with respect to the middle of the sample range. If different samples have different elements equal to zero, there is no problem, zero is also a value.

 
What if we normalise to the maximum/minimum, i.e. just shrink the sample? It turns out that the zero will be the one which was the smallest (not necessarily the first one). By the way, what do you say about changing the range from -0.5 to 0.5. In that case, only the ones that are zero will remain zero. Is it correct?
 
The constant component must also be removed. Consider that the value of the constant component is equal to the first element of the sample, or the middle of the range of the sample. Once the constant components have been removed, determine the maximum range and scale all the samples relative to their zero.
 
Then you have to get that first component right :) The question is how to get it right...
 

So far I am more attracted to the second option - relative to the middle. You can not only use the middle of the range, you can also use the middle value, maybe relative to the middle value will be better.

 
To StatBars something your file is not decompressed please check.
 
rusik1003 писал (а) >>
To StatBars something your file is not decompressed please check.

unzip normally

 
YuraZ писал (а) >>

It's unpacking all right.

Yura, there's a little question.

 
Vinin писал (а) >>

Yura, there's a small question.

ICQ is not working :-(.

mail

yzh

123456789...31
New comment
Reason: