Polyfit and Polyval on Reduced Matrices

Hello!

I'm working with a set of data that needs to have a polyfit and polyval applied to it. However, I initially had to resize the vectors that are being put into the data. Then, I have to plot it.

This is the code that I have. I'm not sure if it's working correctly, as the output numbers for p are nowhere near where I am expecting them to be. It's showing p to be 0.1416, but I'm expecting p to be 50-60...

x and y are both vectors of 2458 x 1 doubles, or I'd attach them here...

figure()

xNew = (linspace(10.5, 16, numel(y)))

p = polyfit(10*log10(xNew),y,1)

f = polyval(p,[10.5 16]);

semilogx(x,y,'o',[10.5 16],f,'-')

Any help is appreciated!

Answer 1

[Torsten]

You are aware that you fit parameters a and b of the function

f(x) = a + 10*b*log10(x)

to your data vector y ?

Did you try

f = polyval(p,10*log10(xNew));
plot(10*log10(xNew),f)
hold on
plot(10*log10(xNew),y,'-')

[AnnieLieseMari]

Thanks for the response! I'm still a little bit confused

I'm not sure where that function the f(x) is...

When I use this code, I get something that still generates p as 0.1416.

figure()

xNew = (linspace(10.5, 16, numel(y)))

p = polyfit(10*log10(xNew),y,1)

f = polyval(p,10*log10(xNew));

plot(10*log10(xNew), f, '-')

hold on

plot(10*log10(xNew), y, '.')

To clarify, I have to limit the x-values to be values from the x vector that are between the values of 10.5 to 16, not the indicies. The values from x that are eliminated then have to be eliminated from y, meaning that x and y would then be the same length.

From there, I have to use a polyfit and polyval. Then, the x-values have to be plotted against the y-values and the linear fit generated by polyfit. But again, it has to only be in that range of x values!

[Torsten]

But you assume that y is a linear function of 10*log10((linspace(10.5, 16, numel(y))). The x-values that are linked with the y-values look quite arbitrary to me.

I'm not sure where that function the f(x) is...

f is p.

plot(10*log10(xNew), f, '-')

hold on

plot(10*log10(xNew), y, '.')

What do you get from the two plot commands ? Is the line you get from the first plot command a good approximation for your data points y ?

[AnnieLieseMari]

So maybe the linspace function is wrong...

What can I do to use only the x values between 10.5 and 16, while extracting the values that correspond from y then?

Yes, the graph is supposed to show the line of best fit to the scatter points that are in x and y.

[Torsten]

What can I do to use only the x values between 10.5 and 16, while extracting the values that correspond from y then?

You define x to be equispaced in 10.5:16 with as many points as the y-vector. So no need to extract them.

I must admit that I don't understand what you are trying to do.