Probability - Linear Regression

Can I get some help for this question?

Below is my attempt

x=[1 3 5 9 11 13];
y=[22 51 127 202 227 248 252];
scatter(x,y)

That is the furthest I can go. Any help would be appretiated. :)

[Stephan]

start with completing your x data - the seven is missing. Then begin to read about least squares curve fitting. To do this you should learn how to code your function in Matlab using a function handle.

[Torsten]

Hint:

log(C/H(t) - 1) = log(A) - B*t

This turns your problem into a linear regression problem.

[Gulfstream Gaming]

@Stephan, Thanks for pointing out the mistake, I just noticed it. Thanks for the links for me to read as well. Appreciate it a lot. See you around!

[Gulfstream Gaming]

@Torsten, thanks for the hint, will appreciate it a lot. :)

See you around

Answer 1

[Star Strider]

I don’t know if you have the Optimization Toolbox or the Statistics and Machine Learning Toolbox, both of which have nonlinear regression functions. One function everyone has is fminsearch (link).

To use it, you will need to optimise a function of the form:

fcn = @(b) norm(y - Hfcn(b,t));

where ‘Hfcn’ is the function you will write for ‘H(t)’, ‘b’ is the parameter vector, and ‘t’ and ‘y’ are the respective data vectors.

I leave the rest to you!

EDIT —

Your ‘t’ vector is also missing a value. It should be:

t=[1 3 5 7 9 11 13];

I discovered that when I actually did the regression.

[Gulfstream Gaming]

@Star Strider, yup, i have both. Thanks for providing me the link. Appreciate it a lot. See you around! I like your profile photo

[Star Strider]

As always, my pleasure.

My polar bear thanks you for the compliment!