plotting leastcurve fit through exceldata

i want to develop a code in which i have to fit a curve using least curve method ,here i have to import the data from excel sheet. from the excel sheet i have to take first 5 points from 1 row 1 column and fit a curve now the again the next 5 points from 2 row 1 column upto n times and fit a curve

i have written for the first 5 points and its working fine. now when i have to slide it to next five points i have given for loop for it and upto that i am getting correct but then i am not getting how to do it further

here is the code for the first 5 points

%% import the data points
function phi = curvefit3(x,y)
[data, ~] = xlsread('accn_rate_ekf01.xlsx'); 

x_vec = data((1:5),1);
y_vec = data((1:5),2);
x0    = x_vec(01);
N = length(x_vec);

%% construction of the least-squares quadratic fit to the data
%% we use the equation y =a0 + a1t + a2t^2
%% numbers a0,a1 and a2 are the unknowns

x_mat = [ ones(N,1) x_vec  x_vec.^2  ]; 
amat  = x_mat'*x_mat;
bmat  = x_mat'*y_vec;
phi   = inv(amat)*bmat;
a0 = phi(1);
a1 = phi(2);
a2 = phi(3); 
[xfit,yfit]   = func(phi,x_vec);
%fun = @a2*(x_vec-x0).^2+a1*(x_vec-x0)+a0;
%yfit = fun(x);
figure;
plot(x_vec,y_vec,'or',xfit ,yfit,'-xg');hold on; grid;
return;
%
%    calculation of function
%
function [xout,yout]   = func(phi,x_vec);
%
x0    = x_vec(01); 
x1    = x_vec(01);
x2    = x_vec(05);
x_array= x1:0.01:x2;
a0 = phi(1);
a1 = phi(2);
a2 = phi(3); 
val= a2*(x_array-x0).^2+a1*(x_array-x0)+a0;
xout= x_array;
yout   = val;
return;

now plzz help me do the same for next five upto n rows

i am not getting how to do it

Answer 1

[sanket neharkar]

i tried and had did this and i am getting expected outcomes ;

can anyone suggest any short method....

function phi = curvefit3(x,y)

clear all;

close all;

[data, ~] = xlsread('filename.xlsx');

x_imp = data((1:n),1); %%reading rows and first column

y_imp = data((1:2n),2); %%reading rows and second column

%figure;

[rows, ~] = size(data);

for i = 1: 1 : n-5 %%loop to read i+1 to i+5 rows

x = data((i+1:i+5),1);

y = data((i+1:i+5),2);

x0 = x(1);

var1 = x - x0;

var2 = var1.*var1;

%% construction of the least-squares quadratic fit to the data

%% we use the equation y =a0 + a1t + a2t^2

%% numbers a0,a1 and a2 are the unknowns

x_mat = [ones(5,1) var1 var2];

amat = x_mat'*x_mat;

bmat = x_mat'*y;

phi = inv(amat)*bmat;

a0 = phi(1);

a1 = phi(2);

a2 = phi(3);

[xfit,yfit] = func(phi,x);

figure;

plot(x,y,'or',xfit,yfit,'-xg');

%hold on;

%hold on

%plot(xfit ,yfit,'-xg'); grid;

end %for i = 1: 1 : n-5

return;

%

% calculation of function

%

function [xout,yout] = func(phi,x);

%

x0 = x(01);

x1 = x(01);

x2 = x(05);

x_array= x1:0.01:x2;

a0 = phi(1);

a1 = phi(2);

a2 = phi(3);

val = a2*(x_array-x0).^2+a1*(x_array-x0)+a0;

xout = x_array;

yout = val;

return;

this is what i was expecting

Answer 2

[dpb]

Pass the data and generalize instead; don't hard code indices into indexing expressions. Also, MATLAB has builtin functions polyfit, polyval use them; they're much more efficient and stable numerically than inverting the design matrix.

% driver
data=readmatrix('accn_rate_ekf01.xlsx');            % xlsread() is deprecated

% function
function [b,yhat,xo,yo] = curvefit3(data,nPts)      % let set number points
  if nargin<2, nPts=5; end                          % default is 5
  nM1=nPts-1;                                       % convenient temp
  nR=size(data,1);                                  % number rows
  i1=1;                                             % starting index
  for i=1:nR-nM1                                    % for how many fits there are
    i2=i1+nM1;                                      % end of section
    x=data(i1:i2,1); y=data(i1:i2,2);               % just a temporary for shorter typing
    b{i}=polyfit(x,y,2);                            % compute coefficients, save as cell array
    yhat{i}=polyval(phi{i},x);                      % save the predicted
    xo{i}=x; yo{i}=y;                               % save the section x,y data as alternate outputs
  end
end

Call the function from your main program; I'd recomend plots be done there, too, not in the computing function

[sanket neharkar]

can you help me with the plots too

i am not getting it