Primer
2008 Nature Publishing Group http:/www.nature.com/naturebiotechnology
What is the expectation maximization algorithm?
Chuong B Do & Serafim Batzoglou
The expectation maximization algorithm arises in many computational biology applications that inv
Prof. Corso, jcorso@buffalo.edu, SUNY Buffalo
Prepared by David Johnson
CSE 455/555 Spring 2013 Quiz 1 of 14
Solutions
Problem 1: Recall (2pts) (Answer in one sentence only.)
What is a decision boundary?
A decision boundary is the region of a problem spac
CSE 455/555 Spring 2013 Quiz 13 of 14
Jason J. Corso, jcorso@buffalo.edu, SUNY Buffalo
Solution by Yingbo Zhou
Name:
ID#:
Section:
455
or
2
555
8
10
Directions The quiz is closed book/notes. You have 10 minutes to complete it; use this paper only.
Problem
Jason J. Corso, jcorso@buffalo.edu, SUNY Buffalo
Solutions by David Johnson
CSE 455/555 Spring 2013 Quiz 4 of 14
Solutions
All problems are worth 2 points.
1. What is one real-world application that last weeks guest lecturers pattern recognition technolog
Jason J. Corso, jcorso@buffalo.edu, SUNY Buffalo
Solutions by David Johnson
CSE 455/555 Spring 2013 Quiz 11 of 14
Solutions
Problem 1: Recall (2pts) (Answer in one sentence only.)
What is the objective criterion of spectral clustering (i.e. what value is
Jason J. Corso, jcorso@buffalo.edu, SUNY Buffalo
Solutions by David Johnson
CSE 455/555 Spring 2013 Quiz 6 of 14
Solutions
Problem 1: Recall (2pts) (Answer in one sentence only.)
What quantity is PCA maximizing during dimension reduction?
PCA maximizes th
CSE 455/555 Spring 2013 Quiz 12 of 14
Jason J. Corso, jcorso@buffalo.edu, SUNY Buffalo
Solutions by Yingbo Zhou
Name:
ID#:
Section:
455
or
555
2
8
10
Directions The quiz is closed book/notes. You have 10 minutes to complete it; use this paper only.
Proble
Jason J. Corso, jcorso@buffalo.edu, SUNY Buffalo
Solutions by Yingbo Zhou
CSE 455/555 Spring 2013 Quiz 9 of 14
Name:
ID#:
Section:
455
or
555
2
2
2
2
2
10
Directions The quiz is closed book/notes. You have 10 minutes to complete it; use this paper only.
A
Jason J. Corso, jcorso@buffalo.edu, SUNY Buffalo
Solutions by Yingbo Zhou
CSE 455/555 Spring 2013 Quiz 7 of 14
Name:
ID#:
Section:
455
or
2
555
8
10
Directions The quiz is closed book/notes. You have 10 minutes to complete it; use this paper only.
Problem
Jason J. Corso, jcorso@buffalo.edu, SUNY Buffalo
Solutions by David Johnson
CSE 455/555 Spring 2013 Quiz 6 of 14
Solutions
Problem 1: Recall (2pts) (Answer in one sentence only.)
What is a support vector?
A support vector is a point that lies (approximate
Jason J. Corso, jcorso@buffalo.edu, SUNY Buffalo
Solutions by Yingbo Zhou
CSE 455/555 Spring 2013 Quiz 8 of 14
Name:
ID#:
Section:
455
or
2
555
8
10
Directions The quiz is closed book/notes. You have 10 minutes to complete it; use this paper only.
Problem
Jason J. Corso, jcorso@buffalo.edu, SUNY Buffalo
Solutions by David Johnson
CSE 455/555 Spring 2013 Quiz 5 of 14
Solutions
Problem 1: Recall (2pts) (Answer in one sentence only.)
Given an unbiased linear discriminant dened by the augmented weight vector a
Solutions provided by David Johnson. See
Code for the solutions to 1-3.
4. SV lists the coordinates of the identied support vectors (for those whove forgotten, the support vectors are
the points nearest the decision boundaryand by extension are the points
L7: Kernel density estimation
Non-parametric density estimation
Histograms
Parzen windows
Smooth kernels
Product kernel density estimation
The nave Bayes classifier
CSCE 666 Pattern Analysis | Ricardo Gutierrez-Osuna | CSE@TAMU
1
Non-parametric density es
Prof. Corso, jcorso@buffalo.edu, SUNY Buffalo
Prepared by David Johnson
CSE 455/555 Spring 2013 Quiz 2 of 14
Solutions
Problem 1: Recall (2pts) (Answer in one sentence only.)
Name one reason one might use a decision tree or forest rather than another type
Spectral Clustering
Braunhofer Matthias (matthias.braunhofer [at] gmail.com)
Strumpohner Juri (juri.strumpohner [at] gmail.com)
Data Warehousing and Data Mining
Free University of Bozen-Bolzano
January 23, 2009
Abstract
Clustering is a popular data mining
Bayesian Decision Theory
Robert Jacobs Department of Brain & Cognitive Sciences University of Rochester
Types of Decisions
Many different types of decision-making situations Single decisions under uncertainty
Ex: Is a visual object an apple or an orange
The Expectation Maximization Algorithm
Frank Dellaert College of Computing, Georgia Institute of Technology Technical Report number GIT-GVU-02-20 February 2002
Abstract This note represents my attempt at explaining the EM algorithm (Hartley, 1958; Dempste
Getting Started
with Matlab
CSE474 Machine Learning
2012 Fall
TA: Yu Liu
yl73@buffalo.edu
Matlab Introduction
Matlab is a program for doing numerical
computation. It was originally designed for
solving linear algebra type problems using
matrices. Its nam
Structure Learning in Bayesian Networks
(mostly Chow-Liu)
Sue Ann Hong
11/15/2007
Chow-Liu
Goal: nd a tree that maximizes the data likelihood
Algorithm
Compute weight I(Xi,Xj) of each (possible) edge (Xi,Xj)
Find a maximum weight spanning tree (MST)
G
CSE 555 Spring 2009 Final Exam
Jason J. Corso
Computer Science and Engineering
University at Buffalo SUNY
jcorso@cse.buffalo.edu
Thursday 7 May 2009, 11:45 AM - 02:45 PM, KNOX 14
Brevity is the soul of wit.
-Shakespeare
The exam is worth 100 points total
CSE 555 Spring 2010 Final Exam
Jason J. Corso
Computer Science and Engineering
SUNY at Buffalo
jcorso@buffalo.edu
Friday 30 April 2010, 8:00 - 11:00, Norton 218
Brevity is the soul of wit.
-Shakespeare
There are 6 questions each worth 20pts; choose 5 of t
CSE 455/555 Spring 2011 Final Exam
Jason J. Corso
Computer Science and Engineering
SUNY at Buffalo
jcorso@buffalo.edu
Date 10 May 2011
Brevity is the soul of wit.
-Shakespeare
Name:
Nickname:
Section:
25
455
or
25
25
25
25
100
555
Nickname is a unique ide
CSE 455/555 Spring 2012 Mid-Term Exam
Brevity is the soul of wit.
-Shakespeare
Jason J. Corso, jcorso@buffalo.edu
Computer Science and Engineering, SUNY at Buffalo
Date 18 Mar 2012
Name:
Nickname:
Section:
25
455
or
25
25
25
100
555
Nickname is a unique i
CSE 455/555 Spring 2012 Final Exam
Jason J. Corso, jcorso@buffalo.edu
Computer Science and Engineering, SUNY at Buffalo
Date 3 May 2012, 11:45 - 14:45
Location Knox 04
Brevity is the soul of wit.
-Shakespeare
Name:
Nickname:
Section:
5
455
or
20
20
20
20
CSE 455/555 Introduction to Pattern Recognition
SUNY at Buffalo
Syllabus for Spring 2013
Last updated: 3 Jan 2013
Instructor:
Jason Corso (UBIT: jcorso)
Course Webpage: http:/www.cse.buffalo.edu/jcorso/t/CSE555 or
http:/www.cse.buffalo.edu/jcorso/t/CSE455
Boosting and AdaBoost
Jason Corso
SUNY at Bualo
J. Corso (SUNY at Bualo)
Boosting and AdaBoost
1 / 62
Introduction
Weve talked loosely about
1
2
Lack of inherent superiority of any one particular classier; and
Some systematic ways for selecting a particul
Spectral Clustering
Jing Gao
SUNY Buffalo
1
Motivation
Complex cluster shapes
K-means performs poorly because it can only find spherical
clusters
Spectral approach
Use similarity graphs to encode local neighborhood information
Data points are vertice
CSE 455/555 Spring 2013 Homework 7: Parametric Techniques
Jason J. Corso
Computer Science and Engineering
SUNY at Buffalo
jcorso@buffalo.edu
This assignment does not need to be submitted and will not be graded, but students are advised to work through the