Ontology and Information Systems Section 1: Basic Concepts

1
Ontology and Information Systems Section 1
Basic Concepts
2
Ontology and Information Systems?

• A popular claim today
• Ontology helps to build better information
  systems.
• This tutorial is meant to show you what this
  could actually mean, from a computer science
  point of view.

3
Outline of this tutorial

• Section 1 Basic Concepts
• Section 2 Ontology, Logic, and Computation
• Section 3 Data Modelling
• Section 4 Ontology-based information systems

4
Outline of Section 1

• A sample Information System
• What the user sees
• What happens inside the system
• More Information Systems
• Alternative ways to design information systems
• Research on Information Systems
• Selected topics
• Summary

5
What is an Information System?
Just our working definition.

• An information system is a software system that
  enables its users to store and access data in a
  database.
• We are all familiar with information systems in
  our everyday life.
• Examples
• The DB reservation system
• An electronic library catalog system
• The web and its search engines

6
What is a database?

• The heart of any information system is its
  database.
• A database is a (structured) collection of data
  stored on a computer.

A database
7
What is a query?

• A major task of an information system is to
  support queries to the database.
• A (database) query is a question intended to be
  answered by a software system based on the
  information stored in a database.

query
answer
A database
A user
8
A sample information system

• Application Domain
• order checking
• Purpose
• check information on customer orders stored in a
  car manufacturers database
• Users
• sales personnel, call center personnel
• Technology
• internet-based software system

9
The User Interface
search parameter
actionbutton
order information
important dates
10
The Users Perspective Define Query
Mr. Meier called twice already. He wants to know
when he will get his new car. I have to tell
him the delivery date next time he calls.

• identify problem

I have to check the status of Mr. Meiers order.
realize how order checking system might help
System does not support the task of finding a
suitable query!
11
The Users Perspective Input Query

• input Mr. Meiers order number

No system support for searching the order number!
submit query
wait
12
The Users Perspective Analyze Results
The interpretation of the systems answer
requires knowledge on the application!

• analyze order information

calculate actual customer delivery date
inform Mr. Meier Youll get your car on 07/15.
13
The Users Perspective Why trust the system?
Business Processes
Event
Action
Responsible

• Enter order information

New order
Sales Person
The Users know where the data comes from.
Order changes
Update order information
Sales Person
Production starts
Update order information
System
14
The users perspective Summary

• The system is used within a problem-solving
  context.
• The system supports only part of the
  problem-solving activities.
• Using the system basically means
• to enter a query
• to analyze the displayed results
• The system is based on a reliable database.

15
The users perspective System Limitations
The system does not ...
Consequences for the user
user must learn in advance whats in the
database contents
help to identify the problem or find a suitable
query
help to input the correct order number
user must know where to get the correct order
number
help to understand the query results
user must learn how to interpret the results or
what can be deduced from it
support more complex queries
use alternative processes or systems to get the
desired information
User must be trained!
16
The users perspective System Limitations
Supported Query Types
Not supported
Show the stored delivery date for a specific
order.
Show available information on a specific order.
Show all orders that have been placed by Mr.
Meier.
Show a list of all customers that have ordered an
A140.
Very limited query facilities
How many A140 have been ordered last month?
...
17
The Systems Perspective Interpret User Input
What happens here, is programmed within the
system and depends on this specific application.

• read user input

construct a complete query
SELECT FROM ORDERWHERE OrderID 0 1 226 60172
hand the query on to the database
18
The Systems Perspective Present Results

• get query results

format query results
OrderID 0 1 226 60172OrderingCustomer
246-4711OrderCar 150 9999 247 ...
output data on the screen
19
The Sample System Summary
User
System

• User and System cooperate to solve a problem.

20
Query Processing The database schema
Name Adress ...
A database schema is a conceptual description of
a databases content.
Typ Colour ...
OrderingCustomer
OrderedVehicle
OrderDate ...
21
Query Processing The database
This is a relational database.
Customer
Table
CustomerID
Customer Name
Customer Adress
Column
Order
OrderID
Ordered Car
Ordering Customer
Vehicle
Row
VehicleID
Typ
Colour
22
Query ProcessingThe database and reality
The real world
Mr. Meier
A Car
The database is intended to mirror the real
world.
Mr. Meiers Order
Customer
CustomerID
Customer Name
Customer Adress
246-4711
Meier
Lennestr., Koblenz
Order
Vehicle
OrderID
Ordered Car
Ordering Customer
VehicleID
Typ
Colour
246-4711
0 1 226 60172
246-4711
246-4711
SLK
Blau
23
Query Processing Why tabular organization?
Besides the relational model there are other
suitable data struc-tures.
Efficient search techniques require a suitable
organi-zation of data.
versus
24
Query Processing A Suitable Query Language
This is relational algebra.

• ... for a tabular database has to support basic
  tasks on tables
• select column
• select row
• compute new table
• sum up values in a column
• etc.

25
Query Processing A Sample Query Language
Definition
This looks like SQL.

• Syntax
• SELECT column_name(s)FROM table_name(s)WHERE
  condition

A standard format for queries is necessary to
ensure unambigous interpretation.
Semantics Perform the described operations on
the database!
A specific program performs the tasks on a given
database.
26
Query Processing An example
Customer
SELECT FROM ORDERWHERE OrderID 01 226
60172
CustomerID
Customer Name
Customer Adress
246-4711
Meier
Lennestr., Koblenz
Order
OrderID
Ordered Car
Ordering Customer
Vehicle
VehicleID
Typ
Colour
0 1 226 60172
246-4711
150 9999 247
246-4711
A-190
Blau
27
Inside the Sample System Architecture
Divide conquer!
User Interface
ApplicationLogic
DatabaseManagement
The inner structure mirrors the main system tasks.

• interpret user input
• format results

• handle input and output

• execute query
• administrate database

System Interfaces
28
Google The User Interface
Query input
!
High recall. Precision?
Results sorted by assumed relevance
29
Google The User Interface
Helps to define more precise queries
30
Google The Database
A heterogenous, distributed database
Specific search strategies required!
... the World Wide Web
31
The Sample System vs. Google
Sample System
Google
homogenous user group
users not known in advance which cannot be
trained systematically
only predefined queries
semi-structured query input
little user support on query input,ranking
supposed to support analysis of results
little user support
huge, heterogenous and distributed database
well organized, relational database of limited
size
data quality unknown
data quality ensured
32
The Sample System vs. Google Note
User
System

• From an ab-stract point of view, Google works
  just like the sample system.

33
The Start System
Natural language query input
34
The Start System
Sometimes, pictures are even more adequat
Natural languageresults
35
Starter and the World Wide Web
Cooperative Answer requires inferencing
36
The Start System

• Understanding the system boundaries is difficult.

Cooperative Answer requires inferencing
37
Natural Language Processing How it works
SyntacticAnalysis
SemanticAnalysis
PragmaticAnalysis
Natural Language Processing requires a lot of
automated reasoning.
Mary
Kiss(Mary,John).
Speakers Intention?
Kiss-es
John
38
The Sample System vs. Start
Sample System
Start
homogenous user group
Users not known in advance which cannot be
trained systematically
only predefined queries
natural language query input
cooperative system that take the users
intention into account
little user support
huge, heterogenous database, including texts,
pictures, etc.
Well organized, relational database of limited
size
data quality unknown
data quality ensured
39
Current research on information systems

• To make information systems ever more usable,
  there is a lot of research going on.
• Topics
• Improving the user interface
• Query input
• Result presentation
• Cooperative Support
• Improving Database technology
• Data structures and search algorithm
• Database integration

40
Database Integration The Task

• The goal is to build up more powerful
  information systems by combining infor-mation
  resources.

41
Database Integration Main Problem

• How can the relationships between data from
  different databases automatically be deduced?

Person
Car
Name Adress ...
Cert.of Oriogin Colour ...
ordered
Car Model
Cert.of Oriogin Colour ...
Contract
Pricing Condition Adress ...
42
Database Integration The Current Solution
c1 getCustomer(db1, customer) c2
getCustomer(db2, person) If c1.name c2.ID
then insert (c1 union c2) into
newDB. ...
In most cases, the integration programs are
handmade.
in many practical applications
43
Ontology and Information Systems

• Current Research concentrates on improving
  databases based on ontological knowledge.
• Selected Topics
• Definition of domain-specific ontologies
• Development of ontologically well-founded data
  models for e.g. biomedical or geographic
  applications
• Definition of quality criteria for data models
  based on ontological priciples
• Implementation of ontology-based software
  components e.g. in web search engines, natural
  language systems or database integration tools

See section 2 4!
44
Summary

• In section 1, you learned the basic terms to talk
  about information systems
• information system general purpose, application
  domain, internal procedures, architecture, design
  decisions
• user interface forms-based query input,
  structured query input, natural language input,
  cooperative user-interface
• database relational database, relational
  algebra, SQL, query processing, database
  management system, database integration
• natural language processing system