Chap4-Part1

# Chap4-Part1 - Linear Models for Classification Discriminant...

This preview shows pages 1–8. Sign up to view the full content.

1 Linear Models for Classification Discriminant Functions Sargur N. Srihari University at Buffalo, State University of New York USA

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

View Full Document
Topics Linear Discriminant Functions – Definition (2-class), Geometry – Generalization to K > 2 classes Methods to learn parameters 1. Least Squares Classification 2. Fisher’s Linear Discriminant 3. Perceptrons 2 Machine Learning Srihari
Discriminant Function Assigns input vector x to one of K classes denoted by C k Restrict attention to linear discriminants – Decision surfaces are hyperplanes First consider K = 2 , and then extend to K > 2 3 Machine Learning Srihari

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

View Full Document
Geometry of Linear Discriminant Functions: Two-class linear disc fn: y (x) = w T x + w 0 w is wt vector and w 0 is bias Assign x to C 1 if y (x) 0 else C 2 Defines boundary as y (x) = 0 • w determines orientation of surface since w T (x A -x B )=0, w is orthogonal to every vector on surface w 0 sets distance of origin to surface w T x/||w|| = w 0 /||w|| y (x) gives signed measure of perpendicular distance r of point x to decision surface , r =y( x )/|| w || With dummy input x 0 =1 and ω =( w 0 ,w) then y (x) = ω T x – passes thru origin in augmented D+1 dim. space
Multiple Classes with 2-class classifiers By using several 2-class classifiers 5 Build a K class discriminant Use K 1 classifiers, each solve a two-class problem Alternative is K(K 1)/2 binary discriminant functions, one for every pair One-versus-the-rest One-versus-one Both result in ambiguous regions of input space Machine Learning Srihari

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

View Full Document
Multiple Classes with K discriminants 6 Consider a single K class discriminant of the form y k (x) = w T k x + w k0 Assign a point x to class C k if y k (x) > y j (x) for all j k Decision boundary between class C k and C j is given by y k (x) = y j (x) This corresponds to D 1 dimensional hyperplane defined by – (w k w j ) T x + ( w k 0 w j0 ) = 0 Same form as the decision boundary for 2-class case w T x + w 0 =0 Decision regions of such a discriminant are always singly connected and convex Proof follows Machine Learning Srihari
Convexity of Decision Regions (Proof) 7 Thus R k is singly-connected and convex ˆ x = λ x A + (1 λ ) x B Consider two points x A and x B both in decision region R k Any point on line connecting x A and x B can be expressed as From linearity of discriminant functions y k (x) = w T k x + w k0 y k ( ˆ x) = λ y k (x A ) + (1 λ ) y k (x B ) Because x A and x B lie inside R k it follows that y k (x A ) > y j (x A ) and y k (x B ) > y j (x B ) for all j k.

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

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

{[ snackBarMessage ]}

### What students are saying

• As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

Kiran Temple University Fox School of Business ‘17, Course Hero Intern

• I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

Dana University of Pennsylvania ‘17, Course Hero Intern

• The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

Jill Tulane University ‘16, Course Hero Intern