Transcript Document

IMSTD:Intelligent Multimedia
System for teaching Databases
By : NAZLIA OMAR
Supervisors: Prof. Paul Mc Kevitt
Dr. Paul Hanna
School of Computing and Mathematical Sciences
Faculty of Informatics
University of Ulster
Intelligent Multimedia System for
Teaching Databases (IMSTD)
 Literature
Review
 Objectives of research + Proposed
work
 Comparison with previous work +
Contribution to the knowledge
 Conclusion
Difficulty in Databases subject
Subject
Very
difficult
Difficult
Introduction to
Databases
-
7.7
76.9
7.7
Entity-Relationship
Modelling
-
48.7
48.7
2.6
12.8
71.8
12.8
2.6
-
71.8
25.6
2.6
2.6
41.0
48.7
5.0
Normalization
The Relational Model
SQL
Easy
Very
easy
Table 1: Percentage of the difficulty of the Databases subject
Objectives of research
To design and implement a transformation tool
To design and implement the components of an ITS
To create a rich, face-to-face learning interaction
through the use of a pedagogical agent
To integrate all of the above components to form
IMSTD
To evaluate students’ and educators’ attitudes
towards this ITS
Literature Review in ITS in
Databases
System
DB_Tutor
(Raguphati and
Schkade, 1992)
Objective
Assist users in database
design
Technique
Using hypertext (in the form of nodes
and links) to present the information on
databases
SQL-Tutor
(Mitrovic, 1998;
Mitrovic and
Ohlsson, 1999)
COLER
(ConstantinoGonzalez and
Suthers, 2000)
ITS in Database
Design (Canavan,
1996)
Supports student learning
SQL
Based on Constraint-Based Modelling
(CBM)
Coach students in entityrelationship modelling in
a collaborative learning
environment
To assists students in
learning Normalization
Based on an architecture for intelligent
collaborative learning (Belvedere)
Menu-based
Literature review in systems that
apply NLP in Databases
System
Aim
Dialogue Tool
(RADD)
(Buchholz et
al., 1995)
To obtain a
skeleton design
of EER model
from designer
DMG (Tjoa
and Berger,
1993)
ANNAPURA
(Eick and
Lockemann,
1985)
To support
designer in
extracting
knowledge from
requirements
specification
To provide a
computerized
environment for
semi-automatic
database design
Type of user
Database
Designer
Techniques used








Dialogue
Syntactic analysis –
ID/LP format
Semantic analysis –
using Jackendoff’s
hypothesis
Heuristics
Attribute Grammar
Pragmatic interpretation
Rules
Heuristics
Dialogue


S-Diagrams
Heuristics

Database
Designer
Database
Designer
Experts of
UoD
Architecture of IMSTD
Agent
User Interface
Transformation
Tool
Domain Model
Tutor Model
Knowledge
Expertise
Teaching goals
Tutoring strategies
Student Model
Student overlay knowledge
Student misconceptions
Prospective Tools

Macromedia Authorware
 Brill’s tagger
 Microsoft Agent
Proposed research work

Step 1 : Read natural language input text
into IMSTD
 Step 2: Part of speech tagging using Brill’s
tagger
Proposed research work
Words tagged
A
department
may
have
several
locations
.
Result
DT
NN
MD
VB
JJ
NNS
.
Meaning
Determiner
Noun, singular mass
Modal
Verb, base form
Adjective
Noun, plural
.
Table 2: Result from Brill’s tagger
Proposed research work

Step 3: Classifying and removing
redundancies and plurals
Sentence
First
Second
Third
Noun
company
departments
department
name
number
employee
department
department
locations
Verb
is
organized
has
manages
Adjective
have
Several
unique
particular
Table 3: Classification of words according to the selected category
Proposed research work

Step 4: Apply heuristics
 Step 5: Refer to history
Word
Entity
Attribute
Relationship
Cardinality
Name
1
10
0
0
Employee
5
0
0
0
Colour
1
5
0
0
Has
0
0
12
0
12
1
4
0
Book
Table 4: An example of the history file
Proposed research work

Step 6: Produce preliminary model
Figure 1: A preliminary model of the scenario
Proposed research work

Step 7: Human intervention
 Step 8: Produce final model
 Step 9: Incorporate into ITS
Comparison with other ITS in
Databases
System
Objective
Technique
DB_Tutor
(Raguphati and
Schkade, 1992)
Assist users in database
design
SQL-Tutor
(Mitrovic,
1998; Mitrovic
and Ohlsson,
1999)
Supports student learning
SQL
Using hypertext
(in the form of
nodes and links)
to present the
information on
databases
Based on
ConstraintBased Modelling
(CBM)
COLER
(ConstantinoGonzalez and
Suthers, 2000)
Coach students in entityrelationship modelling in a
collaborative learning
environment
ITS in Database
Design
(Canavan,
1996)
IMSTD
To assists students in
learning Normalization
To assists students in
learning Data Modelling
Based on an
architecture for
intelligent
collaborative
learning
(Belvedere)
Menu-based
Agent-based
Presence of basic ITS Modules
Domain
Tutoring
Student
module
module
module
Yes
Yes
No
Other
module
NLP
Presence of
agent
No
No
No
Implemented in
the student
module where
the knowledge
is represented in
the form of
constraints
Implemented
under the submodule
Differences
Recognizer
Yes
Yes
CBM
No
No
No
Implemented
under the submodule
Participation
monitor
Coach
module
No
Yes,
limited
feature
Yes
Yes
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
but
Comparison with other systems that
apply NLP in Databases
System
Dialogue Tool
(RADD)
(Buchholz et al.,
1995)
Aim
To obtain a skeleton
design of EER
model from
designer
Database
Designer
Yes
Type of user
Expert Educators
Student
No
No
No
Techniques used
User
Involvement
Language


Yes
German
Yes
German
Yes
No
No
No



Dialogue
Syntactic analysis
– ID/LP format
Semantic analysis
– using
Jackendoff’s
hypothesis
Heuristics
Attribute
Grammar
Pragmatic
interpretation
Rules
Heuristics
Dialogue
Yes
Yes
No
No


S-Diagrams
Heuristics
Yes
English
No
No
Yes
Yes



Brill’s tagger
Heuristics
History file
Yes
English




DMG (Tjoa and
Berger, 1993)
ANNAPURA
(Eick and
Lockemann,
1985)
Transformation
tool (IMSTD)
To support designer
in extracting
knowledge from
requirements
specification
To provide a
computerized
environment for
semi-automatic
database design
To aid students an
educators in
deriving an ER
Model from natural
language text
Contribution to the knowledge

A new technique to transform a natural
language database specification into an ER
model
 The formation of new heuristics
Conclusion

Questionnaire results support the evidence
that Data Modelling is difficult
 Proposed project will contribute to
knowledge
 Worked examples show that the project is
achievable within the time period