{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

Final_solutions - EE 649 Pattern Recognition – Spring...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: EE 649 Pattern Recognition – Spring 2008 Final - Solutions Problem 1. Let C 1 and C 2 denote the classes for the LDA and the nonlinear SVM, respectively. From the plots, we have n = 10, ˆ ǫ n, C 1 = 3 / 10 = 0 . 3, and ˆ ǫ n, C 2 = 1 / 10 = 0 . 1. On the other hand, from the lecture slides, we have V C 1 = d + 1 = 3 and V C 2 = ( d + p- 1 p ) + 1 = ( 3 2 ) + 1 = 4. Therefore, with ξ = 0 . 05, ˆ ǫ n, C 1 + radicalBigg 32 n bracketleftbigg V C 1 log( n + 1)- log parenleftbigg ξ 8 parenrightbiggbracketrightbigg = 0 . 3 + radicalbig 3 . 2(3log 11- log 0 . 00625) ≈ 6 . 6 while ˆ ǫ n, C 2 + radicalBigg 32 n bracketleftbigg V C 2 log( n + 1)- log parenleftbigg ξ 8 parenrightbiggbracketrightbigg = 0 . 1 + radicalbig 3 . 2(4log 11- log 0 . 00625) ≈ 7 . Hence, according to the principle of SRM, one should pick the LDA classifier, despite its having a larger apparent error. Problem 2....
View Full Document

{[ snackBarMessage ]}

Page1 / 2

Final_solutions - EE 649 Pattern Recognition – Spring...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon bookmark
Ask a homework question - tutors are online