Course talk:CPSC312-2017-Linear Regression

From UBC Wiki
Jump to: navigation, search


Thread titleRepliesLast modified
Feedback119:22, 13 October 2017


I like your project idea. What is your objective function? Is it least square error?

You may first work on only doing prediction. That is, given w and X, return y. You need to implement matrix operations.

Then work on fitting a model: given X and Y, return the "best" w with respect to the objective function.

If you still have energy, try a different objective function. But it is optional.

Regards, Rui

00:55, 12 October 2017

Our objective function will be: predict(x_hat, X, ANS) -- so it will take the dataset X and make a linear regression model and use that model to predict x_hat. If we have time, we will add in least squares error.

19:22, 13 October 2017