Data Management in Scientific Domains: Understanding Hierarchies and Concepts - Prof. Bert, Assignments of Computer Science

Information about a university course, ecs 289f topics in scientific data management, focusing on assignments related to rock classification hierarchies and defining concepts using description logic and first-order logic. Students are required to explain the meaning of parent-child relations in hierarchies, provide examples of domains with multiple hierarchies, and discuss knowledge mining using a database instance.

Typology: Assignments

Pre 2010

Uploaded on 07/30/2009

koofers-user-anx
koofers-user-anx šŸ‡ŗšŸ‡ø

5

(1)

9 documents

1 / 1

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
ECS 289F Topics in Scientific Data Management Spring 2006
Dr. Bertram LudĀØascher April 18
Assignment 1
Due: Tuesday, April 25th (in class, after class)
Problem 1
Consider the rock classification hierarchies (Slide 30, Handout 1):1A set of rock-names is hierarchically
classified according to different properties (fabric,chemical-composition,texture).
a) For each of the different hierarchies, explain what the parent-child relation between two nodes
means. What does a node stand for?
b) Give another example of an (ideally scientific) application domain, in which a set of terms (or
names) gives rise to two (or more) different hierarchies. Choose an example, such that a significant
amount of terms occur in more than one hierarchy.
c) Consider a set of terms T={t1, . . . , tn}and two hierarchies isaA/2,isaB/2, i.e., terms Tcan
be classified hierarchically w.r.t. two different organizational aspects Aand B. Now consider a
database instance Iover the schema S={isaA/2,isaB/2}, i.e., a finite set of facts of the form
isaA(tchild, tparent ) and isaB(tchild, tparent ). Discuss how you might ā€œmine new knowledgeā€ about T
and/or Aand B, using I. Use examples from (a) or (b) to illustrate.
Problem 2
Consider the following definitions of concepts:
•AHappyFather is a Man such that all of his children are Rich or Happy.
•AHappyMother is a Woman with at least one child that is Happy.
a) Define HappyFather and HappyMother using description logic (DL) syntax. Use a binary relation
hasChild/2 to model the parent-child relationship.
b) Same as in (a) but use first-order (FO) logic syntax. Hint: Use the translation rules from DL to
FO from class.
Now consider the database schema S={Person(SSN, N ame, Gender),
hasChild(SS N P arent, S SN C hild),
Rich(SSN),Happy(SSN)}
c) Define Man,Woman,HappyFather, and HappyMother as Datalog views over S.
Hint: to express a universal statement (such as ā€œparents having only girlsā€) you might want to
use double negation (e.g. ā€œnot among the parents having no girlsā€).
d) Same as in (c), but use SQL views.
e) Give a database instance Iof Shaving a happy father and mother, and an unhappy father and
unhappy mother. Try to make Ias small as possible.
1The original article will also be available via the class page.

Partial preview of the text

Download Data Management in Scientific Domains: Understanding Hierarchies and Concepts - Prof. Bert and more Assignments Computer Science in PDF only on Docsity!

ECS 289F Topics in Scientific Data Management Spring 2006 Dr. Bertram LudĀØascher April 18

Assignment 1

Due: Tuesday, April 25th (in class, after class)

Problem 1

Consider the rock classification hierarchies (Slide 30, Handout 1):^1 A set of rock-names is hierarchically classified according to different properties (fabric, chemical-composition, texture).

a) For each of the different hierarchies, explain what the parent-child relation between two nodes means. What does a node stand for?

b) Give another example of an (ideally scientific) application domain, in which a set of terms (or names) gives rise to two (or more) different hierarchies. Choose an example, such that a significant amount of terms occur in more than one hierarchy. c) Consider a set of terms T = {t 1 ,... , tn} and two hierarchies isaA/ 2 , isaB/2, i.e., terms T can be classified hierarchically w.r.t. two different organizational aspects A and B. Now consider a database instance I over the schema S = {isaA/ 2 , isaB/ 2 }, i.e., a finite set of facts of the form isaA(tchild, tparent) and isaB(tchild, tparent). Discuss how you might ā€œmine new knowledgeā€ about T and/or A and B, using I. Use examples from (a) or (b) to illustrate.

Problem 2

Consider the following definitions of concepts:

  • A HappyFather is a Man such that all of his children are Rich or Happy.
  • A HappyMother is a Woman with at least one child that is Happy.

a) Define HappyFather and HappyMother using description logic (DL) syntax. Use a binary relation hasChild/2 to model the parent-child relationship. b) Same as in (a) but use first-order (FO) logic syntax. Hint: Use the translation rules from DL to FO from class.

Now consider the database schema S = { Person(SSN, N ame, Gender), hasChild(SSN P arent, SSN Child), Rich(SSN ), Happy(SSN ) }

c) Define Man, Woman, HappyFather, and HappyMother as Datalog views over S. Hint: to express a universal statement (such as ā€œparents having only girlsā€) you might want to use double negation (e.g. ā€œnot among the parents having no girlsā€).

d) Same as in (c), but use SQL views. e) Give a database instance I of S having a happy father and mother, and an unhappy father and unhappy mother. Try to make I as small as possible.

(^1) The original article will also be available via the class page.