CENG 463
Machine Learning
Lecture 15
Neural Networks
Large Feature Space Examples
When feature space (n) is large, logistic regression
is not a good classification algorithm.
Think of a complex classifier
for a twovariable case:
x2
x1
In case of 100 vari
CENG 463
Machine Learning
Lecture 13
Hierarchical Clustering
Why not parametric methods?
A parametric clustering method: Mixture of Gaussians
Clusters are represented as Gaussian distributions, then
the training set is assumed to be a mixture of
Gaussians
Model with One Variable
Example:
House prices
according to area
Price
(1000$)
Size (feet2)
Supervised Learning
Given the right answer for
each example in the data.
Regression Problem
Predict realvalued output
2
Model with One Variable
Training set for
ho
Classification
Logistic Regression is a classification method!
Examples:
Email: Spam / Not Spam?
Brain tumor: Malignant / Benign ?
y cfw_0,1
0: Negative Class (e.g., benign tumor)
1: Positive Class (e.g., malignant tumor)
y cfw_0,1, 2, .
if there are more
Diagnosis for Learning Algorithms
Suppose you have implemented linear regression
to predict housing prices.
However, when you test your hypothesis on a new
set of houses, you find that it makes very large
errors in its predictions. What should you try nex
Unsupervised Learning Reminder
Supervised Learning
Unsupervised Learning
Training set:
cfw_(x(1),y(1), , (x(m),y(m)
Training set:
cfw_x(1), x(2), , x(m)
2
Clustering
We cluster (group) the samples using some features.
Market segment analysis
Social networ
Multiple Variables
By multivariate, we mean there are multiple
variables/features:
Size (feet2) Number of Number of Age of home
bedrooms
floors
(years)
2104
1416
1534
852
5
3
3
2
1
2
2
1
Price ($1000)
46
40
30
36
460
232
315
178
Notation:
= number of feat
Why not parametric methods?
Parametric approaches require knowing the form of the
density.
E.g. With ML estimation in Lecture 2 we assumed that
the underlying function of our data is a Gaussian.
However, in many cases,
 The form is not known
 The form d
Confusion matrix
PredictedClassLabel
Actual Class Label
Actualpositive
Actualnegative
1
0
Predicted positive
1
True positive
False positive
Predicted negative
0
False negative
True negative
There are four possible cases:
For an actually positive sample,

Clustering
Nonparametric (clustering)
No assumptions are made about the underlying densities,
instead we seek a partition of the data into clusters
These methods are typically referred to as clustering
Parametric (mixture models)
These methods model
Gaussian (Normal) Distribution
p(x) = N ( , 2)
1
p( x)
e
2
x 2
2 2
: Mean
: Standard deviation: average absolute difference from the mean.
2: Variance: average squared difference from the mean.
Figure from Introduction to Machine Learning 2ed., E Alpayd
CENG 112 Data Structures
C+ Basics
Mustafa zuysal
[email protected]
February 24, 2017
zmir Institute of Technology
Primitive Types and Operators
Variables
In C+, we have to dene all variables before their rst use.
<type> <identifier>; / or
<type>
CENG 112 Data Structures
Introduction
Mustafa zuysal
[email protected]
February 24, 2017
zmir Institute of Technology
Course Information
Course Contents
This course covers the fundamental data structures such as
linked lists and trees and algorit
REPORT
In the first days of my internship, I have met with managers, engineers and other interns.
They gave information about company and what their organizational cooperation structure is.
In brief, they have such a structure that all of engineers have
DESCRIPTION OF THE COMPANY
Company Name: CBOX Projects Technology
Web: www.cboxprojects.com
Company Location: 1847/10 sok. Cantur Apt. 10/1 Soukkuyu / Yeni Girne/ Izmir / Turkey
Contact person: Hazal Emiroglu
Email: [email protected]
Te
INTRODUCTION
This summer practice was so important for me and I have learn many things in this
period. Apart from learning many things about technically about coding and improving my
current programming, I have also improved my communication skills and te
CHAPTER 4
MOTIVATION
SOME BASIC ASSUMPTIONS
Is that an organization has the right to influence
the behavior of its employees?
1)
2)
3)
4)
Psychological contract
Freedom of choice
No internal or external constraints on behavior
Behaviour is amenable to cha
CENG315 Information Management 2016 Fall Term Tentative Schedule
Week
1
03/10
Course Introduction
2
10/10
Introduction to Information Management in CE
Book2Ch.1
3
17/10
Relational Model
Book1Ch.2 /
Book 3 Ch2
4
24/10
Data Modelling with ERD
Book3Ch.2 /
Contest Winners
Here are the results for this month's contest for best photos in the categories of Child Photos, Flower
Photos, and Scenic Photos. I received hundreds of entries and it was difficult to narrow the entries down
to three in each category. Th
[ Home ]
[ Tips ]
[ Glossary ]
Welcome to CAMshots, a site for people passionate about digital photography. This site has grown out of
decades of photographic experience. I offer advice for both beginners and advanced users. I hope you
enjoy what you find
[ Home ]
[ Tips ]
[ Glossary ]
Tip of the Day
The Rule of Thirds
Good composition often results in good photographs, even of the most mundane subjects. There are
several rules of thumb with respect to employing good composition. One of the most used is th
Who Am I?
Welcome to Dave's Devil Sticks. If you are looking for juggling balls, hacky sacks, pins, unicycles, or
magic hats, you've come to the wrong place; but if you're looking for highquality, handcrafted devil and
flower sticks, this is the site fo
The Basic Stick
The Basic Stick is the perfect stick for beginners. The stick rotates slowly to provide extra time for
performing stick tricks, but is flashy enough to impress your friends.
Patented DuraCoat finish ensures sticks can withstand all weathe
CS 421: COMPUTER NETWORKS
FALL 2015
FINAL
December 28, 2015
150 minutes
Name:
Student No:
Show all your work very clearly. Partial credits will only be given if you carefully state
your answer with a reasonable justification.
Q1
Q2
Q3
TOT
1)
a) (4 pts) Gi
CS 421: COMPUTER NETWORKS
FALL 2014
FINAL
January 6, 2015
150 minutes
Name:
Student No:
Show all your work very clearly. Partial credits will only be given if you carefully state
your answer with a reasonable justification.
Q1
Q2
Q3
TOT
1)
a) (5 pts) Supp
CS 421: COMPUTER NETWORKS
SPRING 2015
FINAL
May 21, 2015
150 minutes
Name:
Student No:
Show all your work very clearly. Partial credits will only be given if you carefully state
your answer with a reasonable justification.
Q1
Q2
Q3
TOT
1)
a) (6 pts) Suppo
CS 421: Computer Networks
FALL 2014
MIDTERM
November 20, 2014
120 minutes
Name:
Student No:
Show all your work very clearly. Partial credits will only be given if you
carefully state your answer with a reasonable justification.
Q1
Q2
Q3
TOT
1)
a) (8 pts)