[Scilab-users] I need your help

[Samuel Gougeon]

Le 19/12/2015 11:17, kumaden ikpambese a écrit :
> Yes. I also wish to predict .i.e ypred and errors associated with the 
> ypred and the y-experimental
> thanks.
>
>
> On Saturday, December 19, 2015 11:10 AM, Samuel Gougeon 
> <sgougeon at free.fr> wrote:
>
> Le 19/12/2015 10:40, kumaden ikpambese a écrit :
>> Please, can you assist me perform multiple regression using reglin 
>> function for following varibles;
>> Input variables;
>>  x_1 = [5.82 5.20 4.80 2.81 2.64];
>> _x _2 = [8.18 6.8 6.2 6.19 7.36];
>> _x _3 = [40 36 32 32 28]; and
>> x_4 = [1.5 1.2 0.87 0.75 0.63].
>> ouput
>> y = [1.5 1.2 0.87 0.75 0.63]
>> Thanks.
> Do you mean that you are searching for the best {a,b,c,d} such that y 
> ~ a.x1 + b.x2 +c.x3 + d.x4 ?


Just do
--> [ x1.' x2.' x3.' x4.'] \ y.'
  ans  =
   - 1.976D-16
     1.308D-16
     0.
     1.

Which means: y = x4   // that is triviously observable on data.
The fitting error is null, since the match between y and x4 is perfect.

S. Gougeon

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.scilab.org/pipermail/users/attachments/20160127/55f0194b/attachment.htm>