plot a curved line of best fit through my data

I have the following discrete points

if i use polyfit with a n = 2 it plots the line shown in the data that is not a good fit of the data. ideally i want a smooth curve between the data that i can then take the equation of to interpolate the values inbetween.

thanks!

Answer 1

[dpb]

Post the data points, but the red line is clearly NOT the least-squares quadratic through the blue points; something is amiss.

If I use estimates of

>> den=[320 500 830]; dep=[0 -20 -110];
>> figure
>> plot(den,dep,'o')
>> b=polyfit(den,dep,2)
b =
   -0.0003    0.1487  -15.1476
>> denfit=linspace(den(1),den(end));yhat=polyval(b,denfit);
>> hold on
>> plot(denfit,yhat,'r-')
>>

A quadratic will fit three points exactly.

To get a smooth curve, you have to evaluate at more than just the three points, but in your plot, the values aren't correct for those points.

Show your work so we can see where the mishap occurred...