Fitting data for each row of a matrix

Hi there,

I have a matrix 8760x8 and I would like to fit the data of each row (8 values) with a vector like this: [1 2 3 4 5 6 7 8]

Is it possible to do it ?

thanks

[Stephen23]

@Nikolas Spiliopoulos: and what is the expected output of this "fitting"?

Answer 1

[dpb]

Guessing the intent Stephen has asked for,

x=1:8;                        % the independent variable vector
nrow=size(data,1);            % rows in data array (8760 but don't use "magic numbers)"
coeff=zeros(nrow,2);          % preallocate for the coefficient array
for ix=1:nrow
  coeff(ix,:)=polyfit(x,data(ix,:),1);  % coefficients of linear fit on each row
end

See

doc polyfit
doc polyval

for details on them. Could also use fit if have Curve Fitting TB and return a fit object containing the coefficients and other info as alternative. If do go this route will have to pass x,row vector as columns.

None of the fitting tools are fully vectorized to do the fit on an array of columns so do need to iterate over the array.

[Nikolas Spiliopoulos]

Hi again,

thanks for the answer. I tried it but I get this error:

Error in (line 10) coeff(i,:)=polyfit(x,annual_temperature_data(i,:),1);

"Subscript indices must either be real positive integers or logicals"

"annual_temperature_data" is my data vector of 8760x8

thanks

[dpb]

Typo --

coeff(ix,:)=polyfit(x,annual_temperature_data(ix,:),1);

i is the Matlab builtin sqrt(-1).

[Daniela Conforti]

Hi, you know how can I use "polyfit" whitout using a for-loop?

I have the same problem described in this question but I'm using a matrix with millions of rows, therefore it requires a lot of time.