I need matlab code to generate smooth curve for data set

x1 =[ 450 600 1500 3000 6000];
y1 =[ 1.8 1.0 0.43 0.3 0.22];

x2 =[ 750 1000 2500 5000 10000];
y2 =[ 3.6 2 0.86 0.6 0.44];

x3=[ 900 1200 3000 6000 12000];
y3 =[ 5.4 3.6 1.29 0.9 0.66];

plot(x1,y1);
hold on;
plot(x2,y2);
hold on;
plot(x3,y3);

title('operation current vs operation time');
xlabel('operation current(A)');
ylabel ('operation time(s)');
grid on;

i used curve fitting tool but cant plot all 3 curves in same grap

Answer 1

[dpb]

There's a way to save the results of the fit session to the command window, but I always have to figure it out and it's more trouble than just doing the fit directly since you know the model you want...

X=[x1;x2;x3].';       % put the data into array by column
Y=[y1;y2;y3].';
f=arrayfun(@(i) fit(X(:,i),Y(:,i),'power2'),1:3,'uni',0);       % fit all three pairs
figure,hAx=axes;hold on                                         % make figure, prepare plot
hL=arrayfun(@(i) plot(f{i},'-',X(:,i),Y(:,i),'x'),1:3,'uni',0); % plot, save line handles

Yields

Above saves the fits in the cell array f, to retrieve each just dereference with the curlies "{}"

>> whos f
  Name      Size            Bytes  Class    Attributes

  f         1x3              3147  cell               

>> f{1}
ans = 
     General model Power2:
     ans(x) = a*x^b+c
     Coefficients (with 95% confidence bounds):
       a =   3.791e+06  (-3.451e+07, 4.21e+07)
       b =      -2.412  (-4.065, -0.7593)
       c =      0.2763  (0.06942, 0.4832)
>> 

See the section on postprocessing linked to from the documentation for fit for all else can be done...

Answer 2

[Thiago Henrique Gomes Lobato]

if you look at the result you will see the fit function for each curve. In your actual image the "f(x)=a*x^b+c" with the coefficients below. You literally just need to plot this function with a given x interval. An example:

x1=400:6200;
y1 = a1*x1.^b1+c1; % Just copy the coefficient values from the result box for each curve
plot(x1,y1)

[dpb]

But, there are only 4 significant digits in the display and one has to manually write the expression besides...not at all pleasant task.

There is a Generate Code option under the File menu that will write the above code (and a lot more that typically isn't needed) that can then use to get the actual coefficients in the model.

There's a Save Session function as well, but all one can do with it is to reload the data back into the tool.