CSE 802, Spring 2012, Homework 2 solutions
DHS, Q7
(a) Decision boundary: x = .
1
dx
= E1
b 1 + (x a1 )/b)2
(a1 )/b
1 dy
= E1
1 + y2
tan1 ( a1 )/b) = E1
2
= a1 +
b
tan(E1 )
Note that classifying a pattern that is actually in 1 as if it where in 2 refer
CSE 802: Homework 1
Due: Jan 17, 2012
Chapter 1 of DHS book introduced an example of a pattern classification system to
separate sea bass from salmon. Along the same lines, consider the following two
classification problems:
(a) A fruit company packages f
Homework 2
CSE 802 - Pattern Recognition and Analysis
Instructor: Dr. Arun Ross
Due Date: February 18, 2014
Note:
1. You are permitted to discuss the following questions with others in the class. However, you must write
up your own solutions to these ques
Homework 2
CSE 802, Spring 2012
Due: Jan 31, 2012
1. Solve the following problems from Chapter 2 of the text book.
7, 12, 13, 27, 31
Please note that the problem numbers provided in the assignment questions correspond to the printed second edition of the
Homework 3
CSE 802, Spring 2012
Due: Feb 14, 2012
1. Let = Probability of Heads of a fair coin. Suppose the coin has
been tossed n times. Let nH be the number of times the coin turned
up heads.
(a) Show that the maximum likelihood estimate of is M LE =
nH
Answer 4
p (x | 1) N(50, 5)
p (x | 2) N(40, 10)
(a) Plot the two class-conditional pdfs in the interval x [10, 75] on the same graph
(b)
(c) and (d)
Answer 6
a) Whitening Transform
Aw
Homework 1
CSE 802: Pattern Recognition and Analysis
Tarang Chugh
A52173951
1. a) The following 4 graphs contain 3 histograms each (1 per class) corresponding to the 4
features in fisher iris dataset. The bin size is selected as 0.2 cm.
Feature 1 (Sepal L
CSE 802: Pattern Recognition
Homework 4
Tarang Chugh
A52175951
Answer 1:
N=100
h = 0.01
h=1
h = 0.1
h = 10
Inference
For smaller values of window width, the probability density is not a smooth
curve. It is because there are not enough number of samples w
Homework 1
CSE 802: Pattern Recognition and Analysis
Instructor: Dr. Arun Ross
Due Date: February 8, 2016
Note: You are permitted to discuss the following questions with others in the class. However, you must write
up your own solutions to these questions
Williams et al. Anim Biotelemetry (2015) 3:45
DOI 10.1186/s40317-015-0077-0
Open Access
RESEARCH
Can accelerometry be used
todistinguish betweenflight types insoaring
birds?
H.J.Williams1*, E.L.C.Shepard1, O.Duriez2 andS.A.Lambertucci3
Abstract
Background
LEARNING TO PLAY CHESS USING
REINFORCEMENT LEARNING WITH
DATABASE GAMES
Henk Mannen
supervisor: dr. Marco Wiering
MASTERS THESIS COGNITIVE ARTIFICIAL INTELLIGENCE
UTRECHT UNIVERSITY
OCTOBER 2003
Of chess it has been said that life is not long enough for
i
Imperial College London
arXiv:1509.01549v1 [cs.AI] 4 Sep 2015
Department of Computing
Girae: Using Deep Reinforcement Learning to Play Chess
by
Matthew Lai
Submitted in partial fullment of the requirements for the MSc Degree in
Advanced Computing of Imper
Homework 4
CSE 802 - Pattern Recognition and Analysis
Instructor: Dr. Arun Ross
Due Date: April 29, 2015
Note: You are permitted to discuss the following questions with others in the class. However, you
must write up your own solutions to these questions.
Homework 5
CSE802, Spring 2012
Due: March 15, 2012
(1) (40 points) Decision Tree
Evaluate the performance of the Decision Tree classifier on iris dataset using 10-fold cross-validation.
Use the MATLAB version of Decision Tree (classregtree).
(a) Report th
Homework 6
CSE 802, Spring 2012
Due: March 27, 2012
For the problems that involve coding, we expect you to give a report on
what you did and the results you got. Please send you codes to the TA
([email protected]) with an email.
1. (35 points) Evaluate the
Homework 7
CSE802, Spring 2012
Due: April 12, 2012
For the problems that involve coding, we expect you to give a report on what you did and the
results you got. Please send you codes to the TA ([email protected]) with an email.
Problem 1
Write a program to
HW3 Sample Solutions
1:
(a) We are given that = prob( Heads ) and number of heads is nH , so the number of
tails is nT = n nH . Let's denote the "Head" outcome as x = 1 , and "Tail" outcome as
x = 0 . So,
n
l ( ) = log likelihood of = ln p ( xi / )
(1)
i
Homework 4
CSE 802, Spring 2012
Due: Feb 27, 2012
1.
(a) Compute the 2 x 2 scatter matrix S for this data
n
S = (x j x )(x j x ) , where x j is the j-th sample (row) and x is the mean vector that is
T
j =1
calculated by using all the samples. Under these
Homework 5
CSE 802, Spring 2012
Due: March 15, 2012
1.
% x is the data matrix, i.e. 150x4
% y is a cell structue of labels (string values)
for i=1:10
ind=randperm(150);
x2=x(ind(1:135),:);
ys2=ys(ind(1:135);
x3=x(ind(136:end),:);
ys3=ys(ind(136:end);
t2=c
CSE802
04/09/2012
HW 6 Solutions
Thanks to Mehrdad Mehdavi who has shared his solutions and template with us. This
answer key is a modied version of his assignment.
Problem 1
(a)
One versus Rest
*
accuracy =
1.00
1.00
0.933
0.86
1.00
1.00
0.86
1.00
0.86
0
Name: Practice Quiz - 1
CSE 802 - Pattern Recognition and Analysis
Posted on: February 11, 2014
1. [6 points] Briey explain the following terms with the appropriate formulae: (a) Bayes Risk;
(b) Bayes Rule.
2. Consider a two-category one-feature classicat
Homework 1
CSE 802: Pattern Recognition and Analysis
Instructor: Dr. Arun Ross
Due Date: Jan 28, 2014
Note: You are permitted to discuss the following questions with others in the class. However, you must write
up your own solutions to these questions. An