Pattern Recognition Quiz 1 with 3 Problems - 2005 | CS 791V, Quizzes of Computer Science

Material Type: Quiz; Professor: Ross; Class: ADTP:Statistcl Pattrn Recogntn; Subject: Computer Science; University: West Virginia University; Term: Fall 2007;

Typology: Quizzes

Pre 2010

Uploaded on 07/31/2009

koofers-user-n3f
koofers-user-n3f 🇺🇸

10 documents

1 / 1

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
Name: ———————-
Practice Quiz - 1
CS 591U/791V - Pattern Recognition
Posted on: September 27, 2005
Note:
Univariate normal density: N(µ, σ2)=1
2πσ eh1
2(xµ
σ)2i.
1. [3 points] List the 3 general forms of learning that can be used by a pattern recognition system.
2. Consider a two-category problem where the true state of nature is represented by the variables ω1
and ω2. Let αi(i= 1,2) be the action taken on observing a particular input x. Let λij indicate the
loss incurred when taking action αigiven that the true category is ωj.
(a) [2 points] Write down expressions for the conditional-risks R(α1|x)and R(α2|x).
(b) [4 points] Derive an expression for the likelihood ratio p(x|ω1)/p(x|ω2).
3. [6 points] Let p(x|ωi)N(µi, σ2)for a two-category one-dimensional problem with P(ω1) =
P(ω2) = 1/2. Show that the minimum probability of error is given by,
Pe=1
2πZ
a
eu2
2du,
where a=|µ2µ1|/2σ.
1

Partial preview of the text

Download Pattern Recognition Quiz 1 with 3 Problems - 2005 | CS 791V and more Quizzes Computer Science in PDF only on Docsity!

Name: ———————-

Practice Quiz - 1

CS 591U/791V - Pattern Recognition

Posted on: September 27, 2005

Note: Univariate normal density: N(μ, σ^2 ) = √ 21 πσ e

h − 12 (x− σμ ) 2 i .

  1. [3 points] List the 3 general forms of learning that can be used by a pattern recognition system.
  2. Consider a two-category problem where the true state of nature is represented by the variables ω 1 and ω 2. Let αi (i = 1, 2 ) be the action taken on observing a particular input x. Let λij indicate the loss incurred when taking action αi given that the true category is ωj. (a) [2 points] Write down expressions for the conditional-risks R(α 1 | x) and R(α 2 | x). (b) [4 points] Derive an expression for the likelihood ratio p(x | ω 1 )/p(x | ω 2 ).
  3. [6 points] Let p(x | ωi) ∼ N (μi, σ^2 ) for a two-category one-dimensional problem with P (ω 1 ) = P (ω 2 ) = 1/ 2. Show that the minimum probability of error is given by, Pe = √^12 π

a^ e

− 2 u (^2) du,

where a = | μ 2 − μ 1 | / 2 σ.