LFD Book Forum (http://book.caltech.edu/bookforum/index.php)
-   Chapter 1 - The Learning Problem (http://book.caltech.edu/bookforum/forumdisplay.php?f=108)
-   -   How to draw the separating line in 2D input (http://book.caltech.edu/bookforum/showthread.php?t=4422)

 Ahmed.Elkayesh 09-25-2013 08:20 PM

How to draw the separating line in 2D input

This is my first post to the forum and I would like to thank the book authors for the availability online and for choosing to make the book price low.

:clueless:
I have a question and I hope someone will help me answer it

In a 2D input space, we have the components in the W vector, w0 for the bias, and w1, w2 for the 2 dimensions of the input.

If the learning algorithm successfully found the W that could classify all input correctly, how is it possible to draw the separation line(or vector) using the 3 components of the W vector?

How could we map w0,w1,w2 to the Line equation y = ax+ b.

Thanks

Ahmed

 yaser 09-27-2013 10:18 PM

Re: How to draw the separating line in 2D input

Quote:
 Originally Posted by Ahmed.Elkayesh (Post 11510) This is my first post to the forum and I would like to thank the book authors for the availability online and for choosing to make the book price low. :clueless: I have a question and I hope someone will help me answer it In a 2D input space, we have the components in the W vector, w0 for the bias, and w1, w2 for the 2 dimensions of the input. If the learning algorithm successfully found the W that could classify all input correctly, how is it possible to draw the separation line(or vector) using the 3 components of the W vector? How could we map w0,w1,w2 to the Line equation y = ax+ b. Thanks Ahmed
The line would be which expands as with and being the coordinates in the conventional formulation. As you can see, you need to diivide out by to get the coefficient of which is to be 1, hence the discrepancy between the 3 parameters in one formulation and the 2 parameters in the other formulation.

 All times are GMT -7. The time now is 02:43 AM.