CS464 Introduction to Machine Learning
Fall 2009
Homework 1 – Concept Learning
Due Date:
October 14, 2009
Assume that the following training examples are given:
Ex
Attrb1
Attrb2
Attrb3
Attrb4
Attrb5
TargetAttrb
1
a
b
b
b
a
yes
2
a
b
b
a
a
yes
3
b
c
a
a
b
no
4
c
b
b
b
b
yes
where possible values for attributes area as follows:
Possible values for Attrb1: a, b, c
Possible values for Attrb2: a, b, c, d
Possible values for Attrb3: a, b
Possible values for Attrb4: a, b
Possible values for Attrb5: a, b, c
Possible values for TargetAttrb: yes, no
Assume that each hypothesis is described by a conjunction of constraints on the attributes.
Q1) Answer the following short question about the hypothesis space.
i)
How many instances are there in the instance space X?
ii)
How many semantically distinct hypotheses are there in the hypothesis space H?
iii)
What is the set of instances that the hypothesis <?,a,a,?,a> represents?
iv)
What is the set of hypotheses that is represented by the following version space:
S: <a,?,b,?,b>
G: { <?,?,b,?,b>, <a,?,b,?,?> }
Q2)
i)
Find the target function (the hypothesis) that is learned by FINDS algorithm by giving the trace of FINDS
This is the end of the preview.
Sign up
to
access the rest of the document.
 Fall '08
 Demir
 Machine Learning, possible values, Mücahid Kutlu

Click to edit the document details