help-octave
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: need help with polyfit and polyval


From: Dmitri A. Sergatskov
Subject: Re: need help with polyfit and polyval
Date: Mon, 6 May 2019 09:54:11 -0500

On Mon, May 6, 2019 at 9:50 AM Dmitri A. Sergatskov
<address@hidden> wrote:
>
>
>
> On Mon, May 6, 2019 at 9:31 AM Doug Stewart <address@hidden> wrote:
>>
>> I am trying to convert some data from wikipedia to octave and I ran into 
>> this problem.
>>
>> here is the code
>> **********************************************
>> data = [    4.0000    3.1000
>>     5.0000    4.7400
>>     6.0000    6.1300
>>     7.0000    7.2600
>>     8.0000    8.1000
>>     9.0000    8.1400
>>    10.0000    8.7400
>>    11.0000    8.7700
>>    12.0000    9.1300
>>    13.0000    9.1400
>>    14.0000    9.2600  ];
>>   x = data(:,1);
>>   y = data(:,2);
>>   [P, S, MU] = polyfit (x, y, 1)
>> y3=S.yf
>> y2=polyval(P,x)
>>
>> plot(x,y,'x',x,y,x,y2,x,y3)
>>
>> **********************************************
>>
>>
>> poly fit  should return the polynomial in P
>> But when i use it in polyval it is obviously wrong.
>> but polyfit also returns the new calculated y values in S.yf.
>> S.yf look correct ut P seems to be wrong.
>> I get
>> P =   1.8410   7.5009
>> which is      y=x*1.841 + 7.5009
>> and this is obviously wrong.
>
>
> Yes. It looks like
> [P, S, MU] = polyfit (x, y, 1) form has a problem.
> if you just do p = polyfit (x, y, 1)
> you get a correct p:
> P=polyfit (x,y,1)
> P =
>
>    0.55509   2.50509
>
> octave:12> y2=polyval(P,x)
> y2 =
>
>     4.7255
>     5.2805
>     5.8356
>     6.3907
>     6.9458
>     7.5009
>     8.0560
>     8.6111
>     9.1662
>     9.7213
>    10.2764
>

In fact it looks like a feature, from the help:

When the third output, MU, is present the coefficients, P, are
     associated with a polynomial in

     'XHAT = (X - MU(1)) / MU(2)'
     where MU(1) = mean (X), and MU(2) = std (X).


I think text in help can be improved

>
>
> Dmitri.
> --
>

Dmitri.



reply via email to

[Prev in Thread] Current Thread [Next in Thread]