Thread: Regression

I have 20 points in the form of x and y.

And i am asked to

Assume that there may be the following functional relationships:

y

= w0 + w1*x, or y = w0 + w1*(x^2)

Perform least squares regression for both and determine what the

best

functional form is.
(

Hint: use the first half of the data as training set and the second half

as test set.)

What am i meant to do, i have no idea where to begin.

Thanks Adam

I'm a computer scientist so I can't tell you about all the ins and outs but as you don't know where to even start maybe some of this will prompt you in the right direction:

You need to do linear regression analysis on the first 10 items of your data set.
Simple linear regression - Wikipedia, the free encyclopedia
This will give you the best form for y = w0 + w1*x

I think you can also do linear regression again to find w0 and w1 for y = w0 + w1*x^2 if you square all your y variables first.

Then you'll need to plug the x values of your test set into the forms and find and square the difference between the last 10 items and the obtained y. The one with the least squared error over these last 10 data will be the one you can conclude is the best form.

I'm not a mathematician. I've just read a bit around the subject before. So take what I've said with a grain of salt if it didn't start the cogs turning.

Hope this helps.
-Matthew