x = 0:0.1:1;
y = sin(x); % your data
p = polyfit(x(3:4),y(x>=x(3) & x<=x(4)),1); % fitting coefficients
yf = p(2) + p(1)*x(3:4); % curve fit
plot(x,y,x(3:4),yf,'r*')
Of course, you want to make sure that the fitting is done over a "small" region (just as an example, between 3rd and 4th data points, as shown above) to get a "good" approximation of the in-between vals.
