ECS 271, Machine Learning: Homework #1
Due: 8 April 2004

Hand-in your assignments in the class.

Instructor: Prof. Rao Vemuri, rvemuri@ucdavis.edu

1. (10 points) Please look at the list of potential projects given elsewhere on this web site. Either pick one of those or come up with a suggestion of your own. Prepare your proposal along the lines described on the "Projects" page of this site. Your project has to do with one of the ML topics discussed in the text. You can pick any application domain.

When I grade your term project, I will consider both the quality of your work and the inherent difficulty (or potential) of the problem. The challenge is to pick a small enough topic that you can do within 9 weeks and a topic about which you lnow something.

2. (20 points) Read Chapter 1 and do Problem 1.3, page 18 of the text book

3. (20 points) You are given the following "user profile" problem for web browsing.

Domain	Platform	Browser	Day	Screen	Continent	Click?
edu	Mac	Mozilla	Mon	XVGA	America	yes
com	Mac	NetCom	Tue	XVGA	America	yes
com	PC	IE	Sat	VGA	Eur.	no
org	Unix	Net5	Wed	XVGA	America	yes

The attributes of the instances x  (x is an element of the instance space X) have the following possibiliities
- Domain: edy, com, org
-Platform: Mac, PC, Unix
-Browser: Mozilla, Netscape 5, Netscape Communicator, Microsoft IE
-Day: Monday - Sunday
-Screen: VGA, XVGA
-Continent: America, Europe, Africa, Asia, Australia

• What is the cardinality of the instance space X? Why?
• How many syntactically disticnt hypotheses are there in the hypothesis space H? Why?
• How many semantically distinct hypotheses are therein H? Why?

SKIP PROBLEM 3 BELOW  FOR NOW. We will do it later
3. (20 points) Suppose you are a doctor treating a patient who occasionally suffers from an allergic reaction. Your intuition tells you that the allergy is a direct consequence of certain foods the patient eat, place where it is eaten, time of day, day of week and the amount spent on food. So we gathered some data and the data is summarized in the table below.

Number	Restaurant	Meal	Day	Cost	Reaction
1	Sam's	breakfast	Friday	cheap	yes
2	Hilton	lunch	Friday	expensive	no
3	Sam's	lunch	Saturday	cheap	yes
4	Denny's	breakfast	Sunday	cheap	no
5	Sam's	breakfast	Sunday	expensive	no

Use the notation [?, ?, ?, ?] for the most general model, where the ? indicates a don't care condition (i.e., any attribute value in the corresponding position is allowed.)

What is the hypothesis that fits this data using (a) Find-S algorithm described in Section 2.4 and the Version Space method described in the text. Do not simply give me an answer; show me a trace of the steps that led you to the result. Add any commentary to clarify the steps.