Flexible querying of relational databases fuzzy set based approach 27-11

Flexible Querying of Relational Databases: Fuzzy Set Based Approach
By
Adel A. Sabour, Ahmed M. Gadallah, and Hesham A. Hefny
Flexible Querying of Relational Databases:
Fuzzy Set BasedApproach
Advanced Machine Learning Technologies and Applications
Communications in Computer and Information Science Volume 488, 2014, pp 446-455
6:35:14 PM

Agenda
Presentation Outline:
1)Introduction
2)Crisp SQL vs. Fuzzy SQL
3)Motivation
4)Research Problem
5)The Objective
6)Previous work
7)The Proposed Fuzzy Query Approach
8)Fuzzy Queries (FQ) Tool
9)Results
10)Summary
6:35:14 PM

Introduction
Structured Query Language (SQL) is a very essential
language for querying relational databases. It manipulates
and retrieves data which is crisp and precise by nature. In
contrary, it is unable to respond to human-like queries
which are uncertain, imprecise and vague in nature.
Almost, human queries have a lot of vagueness and
ambiguity due to using his/her subjective linguistic words.
Introduction:
6:35:14 PM

Introduction
For example, high salaries have different definitions
that depend on each person searching for them.
Consider the query “retrieve the names and addresses
of the university students who have height around the
ideal height for a handball player”. This query cannot be
expressed and manipulated directly by a traditional SQL
statement.
Introduction:
6:35:14 PM

Introduction
In contrary, it can be expressed and manipulated easily
through a fuzzy query statement in a very flexible and
human-like manner based on the ideas of fuzzy set theory.
Human-like queries depend essentially on manipulation
of linguistic values such as: short, tall, hot and calls human
concepts.
Introduction:
6:35:14 PM

Introduction
standard SQL adopts classical Boolean expressions with
crisp logical connectors which are too rigid and very limited
in manipulation of linguistic values or linguistic modifiers
such as: about, nearly, fairly, etc. which can be represented
easily based on the concept of fuzzy set.
Introduction:
6:35:14 PM

Crisp SQL vs. Fuzzy SQL
Example: “Give me the names of the young age and high salary
employees”. This query is modelled in SQL language as follows:
SELECT Emp_name, Emp_age, Emp_salary
FROM Employee
WHERE (Emp_age between 22 and 30)
and (Emp_salary>8)
Basic Concepts of fuzzy set theory
6:35:14 PM

The attribute age, presented in next figure, has the linguistic labels
Young, Adult and Old, defined on the trapezoidal possibility
distributions as following:
YOUNG (18, 22, 30, 35)
ADULT (25, 32, 45, 50)
OLD (50, 55, 62, 70)
An approximate value has a margin of 5.
Figure. fuzzy membership function for AGE attribute.6:35:14 PM

The attribute SALARY, presented in next fig., has the linguistic labels
Low, Medium and High, defined on the trapezoidal possibility
distributions as following:
LOW (1, 1.5, 2.5, 4)
MEDIUM (3, 4.5, 6.5, 8)
HIGH (6, 8, 10, 12)
An approximate value has a margin of 1.5 L.
Figure. fuzzy membership function for Salary attribute6:35:14 PM

Query syntax after those Settings will be written as follows in our
proposed system:
SELECT Emp_name, Emp_age, Emp_salary
FROM Employee
WHERE Age.Young(Emp_age )
and Salary.High(Emp_salary)
6:35:14 PM

Motivation
Most of previously proposed fuzzy query approaches have a
disadvantage that they interpret any fuzzy query statement into a
crisp query statement then evaluate the resulted tuples to compute
their matching degrees to the fuzzy query.
This proposed approach makes it easy to use fuzzy linguistic
values in all clauses of a select statement, not “where” clause only.
The added value of this proposed approach is to accelerate the
execution of fuzzy query statements.
Motivation
6:35:14 PM

Research Problem
Most of previously developed fuzzy SQL architectures have set of
drawbacks:
1) Have the inability for dealing with multi subjective views when
interacting with multiusers. Instead, preferences are used to help in
reducing the amount of information returned in response to user
queries.
2) Lack of a standardized format in writing statements that is each
tool has its own syntax. On the other hand, most of the proposed
approaches depend mainly on using a time consuming
parser/translator to check and convert fuzzy queries to crisp SQL
queries.
Research Problem
6:35:14 PM

Research Problem
3) Most of proposed approaches depend mainly on using a
time consuming parser/translator to check and convert
fuzzy queries to crisp SQL queries.
4) Some approaches process nested and correlated fuzzy
queries inefficiently.
Research Problem
6:35:14 PM

Research Problem
5) Most of approaches define a Meta base named Fuzzy
Meta-knowledge Base (FMB) that includes a set of tables to
store all necessary information to describe and manipulate
fuzzy attributes and terms values.
This slows down the process of querying because each
generated fuzzy query must be analyzed and translated into
a standard SQL statement respecting the definitions of all
used fuzzy terms or operators stored in FMB.
Research Problem
6:35:14 PM

The Objective
The Objective
The main objective of the proposed fuzzy query
approach is to overcome the above disadvantages by
evaluating each tuple directly through the use of stored
database objects namely packages, procedures and
functions. Consequently, the response time of executing
a fuzzy query statement will be reduced.
6:35:14 PM

Human-Like Data-Sensitive fuzzy-Based Query Approach for Relational
Databases
Previous work
Previous work
FSQL proposed by Galindo [10] has been conceived with a mild coupling
architecture as shown in figure. FSQL is integrated through stored procedures and
functions.
Figure 4. FSQL implementation architecture (FRDB) by Galindo [10]

Databases
Previous work
Previous work
FSQL proposed by Galindo [10]
The drawbacks:
1. The proposed approach depend mainly on using a time consuming
parser/translator to check and convert fuzzy queries to crisp SQL queries.
2. The approach define a Meta base named Fuzzy Meta-knowledge Base (FMB)
that includes a set of tables to store all necessary information to describe and
manipulate fuzzy attributes and terms.
Such FMB must be accessed each time a fuzzy query statement is
generated. This operation is essential in order to obtain the definitions
of each used fuzzy term in order to complete the processing of the
generated fuzzy query statement which is a time consuming.
3. No querying of relational databases using non-stationary or data sensitive
fuzzy values.

Databases
Previous work
Previous work
Also, another approach for fuzzy query called SQLf is proposed by Bosc in [11]. As
shown in its architecture in Figure.
Figure 5. SQLf implementation architecture by Bosc & Pivert [11].

Databases
Previous work
Previous work
FSQL proposed architecture by Bosc & Pivert [11].
The drawbacks:
1. The proposed approach depend mainly on using a time consuming
parser/translator to check and convert fuzzy queries to crisp SQL
queries.
2. The approach define a Meta base named Fuzzy Meta-knowledge
Base (FMB)
3. No querying of relational databases using non-stationary or data
sensitive fuzzy values.
4. The lack of a standardized format in writing statements , each
tool has its own syntax.

The Proposed Fuzzy Query Approach
6:35:14 PM

The proposed fuzzy query approach aims mainly to
support human-like queries which almost contain
linguistic terms and fuzzy connectors. Such linguistic
terms include linguistic variables, linguistic values, fuzzy
hedges and fuzzy numbers. The proposed approach
enhances the definition and manipulation of such
linguistic terms.
6:35:14 PM

Each defined linguistic term will be stored as a database
object namely stored function within a related stored
package. In consequence, a user can use any of his
defined linguistic terms in any clause in a select
statement as using a user defined stored function.
Accordingly, the defined linguistic terms can be used
easily in even nested, complex and correlated query
statements without the need for complex procedures to
execute the query statement.
6:35:14 PM

A Fuzzy Query Tool based on the proposed approach
A Fuzzy Queries Tool
Register/Login to FQ
At the first time the user aims to interact with the tool,
he/she must create an account with a user name and a
password. After that, the user can define his/her own
linguistic terms via the tool GUI. Consequently, each
created stored database object within the user session
becomes part of the user fuzzy profile. Also, the user
can generate a fuzzy query statement using his/her
predefined linguistic terms and execute it.
6:35:14 PM

Linguistic Variables Service
This service allows creating, modifying or deleting a linguistic
variable/labels that has a set of linguistic values that will be
defined over it. The user enters just the name of the linguistic
variable. For example, a linguistic variable may be a computer
grades, temperature degree, height or salary as shown in next
figure.
6:35:14 PM

Linguistic Values Services
This service is responsible for defining a set of linguistic
values over a specific linguistic variable. Through a graphical
user-friendly representation, the user can check the linguistic
values definitions and its overlapping for a specific linguistic
variable. The graphical representation will help the user to
make the effective modifications in the selected membership
function to satisfy the meaning of the linguistic value.
6:35:14 PM

Linguistic Values Services
The user can draw these
values (Control Points)
only, or add to the values
set by or delete the
previous drawing. Then
add these values to the
Linguistic Value to add
other Linguistic Value for
the same Linguistic
Variable. And so on, until
the user saves all the
values6:35:14 PM

Fuzzy SQL Query Service
This service enables the user to generate fuzzy query statements
using the syntax of traditional SQL language. After the execution
of the generated fuzzy query statement, the result it displayed as
shown in next Fig.
The next query statement is generated to obtain each employee
first name, salary and a matching degree specifying how much
such salary is high
6:35:14 PM

Fuzzy SQL Query Service
6:35:14 PM

Results
Results
Using a fuzzy expression in where clause
Passing a fuzzy expression as an argument to a function
6:35:14 PM

Results
Results
complex queries including joining of two or more tables are
supported. The resulted tuples can be sorted using their matching
degree to the specified fuzzy criteria in the generated fuzzy query
statement.
This query statement
aims to retrieve each
student Id, student
name, physics grade
and a matching
degree specifies how
much such student if
failed in physics
6:35:14 PM

Results
Results
Creating a database view based of fuzzy expressions
An example of a correlated Fuzzy query statement
6:35:14 PM

Results
Results
In order to enhance the previous query statement to show each
employee name, salary, the average salary of his/her department
and a matching degree representing how much the employee
salary is closed to such average, the fuzzy query statement shown
in next fig. is used. The result of such a query statement shows
that the more closed salary to the average, the higher the
matching degree
6:35:14 PM

Summary
Summary
the proposed approach depends mainly on storing all used fuzzy
terms as database objects in an organized fashion. Such stored
database objects resembles the user fuzzy profile.
the fuzzy query syntax is identical to the standard SQL language
due to the use of stored database objects for representing the
fuzzy terms. In other words, any generated fuzzy query statement
agrees with the standard SQL language.
6:35:14 PM

Summary
Summary
Consequently, a fuzzy query statement can be generated directly
as a traditional SQL statement using the fuzzy terms predefined as
database objects. In consequent, there is no need for an analyzer
or a translator to covert a generated fuzzy query into an
executable query statement.
Also, there is no need for a Fuzzy Meta-Knowledgebase that
existed in most of previously proposed approaches for fuzzy
queries. Hence, not only the complexity of allowing a fuzzy query
is reduced but also the response time of executing such query.
6:35:14 PM

Summary
Summary
The proposed approach allows simple, complex, nested and
correlated fuzzy query statements in a human-like fashion which
almost contain linguistic terms and fuzzy connectors. Also,
because linguistic terms are represented as a set of user defined
stored functions; the proposed approach allows writing any
simple, nested, correlated and complex fuzzy query statements as
traditional select statements in a very flexible manner.
6:35:14 PM

(Questions)
Flexible Querying of Relational Databases :
Fuzzy Set Based Approach
Speaker: Adel Sabour
Email: adelsabour@gmail.com
Phone: 01120200131
Advanced Machine Learning Technologies and Applications
Communications in Computer and Information Science Volume 488, 2014, pp 446-455
6:35:14 PM

References
1) Mishra, J.: Fuzzy Query Processing. International Journal of Research and Reviews in
Next Generation Networks 1(1) (March 2011)
2) Grissa, A., Ben Hassine, M.: New Architecture of Fuzzy Database Management Systems.
The International Arab Journal of Information Technology 6(3) (July 2009)
3) Qi, Y., et al.: Efficient Processing of Nested Fuzzy SQL Queries in a Fuzzy Database. IEEE
Transactions on Knowledge and Data Engineering 13(6) (November/December 2001)
4) Abbaci, K., Lemos, F., Hadjali, A., Grigori, D., Liétard, L., Rocacher, D., Bouzeghoub, M.:
Selecting and Ranking Business Processes with Preferences: An Approach Based on
Fuzzy Sets. In: Meersman, R., Dillon, T., Herrero, P., Kumar, A., Reichert, M., Qing, L.,
Ooi, B.-C., Damiani, E., Schmidt, D.C., White, J., Hauswirth, M., Hitzler, P., Mohania, M.,
et al. (eds.) OTM 2011, Part I. LNCS, vol. 7044, pp. 38–55. Springer, Heidelberg (2011)
5) Singh, K., et al.: Study of Imperfect Information Representation and FSQL processing.
In- ternational Journal of Scientific & Engineering Research 3(5) (May 2012)
6) Garg, A., Rishi, R.: Querying Capability Enhancement in Database Using Fuzzy Logic.
Global Journal of Computer Science and Technology 12(6) (Version 1.0 March 2012)
7) Wahidin, I.: Fuzzy Control (2007)
References:
6:35:14 PM

References
8) Gadallah, A., Hefny, H.: An Efficient Database Query Processing Tool Based on Fuzzy
Logic. In: The 37th Annual Conference on Statistics and Computer Science (2002)
9) Shawky, A., et al.: FRDBM: A Tool For Building Fuzzy Relational Databases. The Egyp- tian
Computer Journal (2006)
10) Galindo, J., et al.: Handbook of Research on Fuzzy Information Processing in Databases.
Chapter XI FSQL and SQLf: Towards a Standard in Fuzzy Databases (2008)
11) Bosc, P., Pivert, O.: SQLf Query Functionality on Top of a Regular Relational Database
Man- agement. In: Pons, O., Vila, M.A., Kacprzyk, J. (eds.) Proceedings of Knowledge
Manage- ment in Fuzzy Databases. STUDFUZZ, vol. 39, pp. 171–190. Springer, Heidelberg
(2000)
12) Moore, S.: Oracle Database PL/SQL Language Reference, 12c, p. 1 (2014)
13) Elmasri, R., Navathe, S.B.: Fundamental of Database Systems (2014)
References:
6:35:14 PM

Flexible querying of relational databases fuzzy set based approach 27-11

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Flexible querying of relational databases fuzzy set based approach 27-11 (20)

Recently uploaded (20)

Flexible querying of relational databases fuzzy set based approach 27-11