•
•
!
$
% !
"
$
#
% &$
!
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
2
•
$ % $
• )%
% '
(
%
% $
•
' – /
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
%
%
! *%
+,,' ' ' -
% '
- !,.
%
3
• )
$
! 0 1
$ • )%
%
– $
! % %
% $
!
"
2% '
%
!# –
( !% %
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
% '
! (
% $
4
•
$ – –)
% !
$
% 0
$ 1
%$
• – $
%
! %
! $
$
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
5
Parameters Synthetic Schema (S1)
Schema Generator Synthetic Schema Extractor
Synthetic Schema (S2)
Synthetic Schema Modifier n
... Synthetic Schema (Sj)
Base Schema (S)
Domain Ontology
• • • 3
% +
% %
! 1 $
%
1 %
1
!
0
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
6
Base Schema: IMDb
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
7
! Empty Synthetic Schema
Candidate Elements = E
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
1-size Synthetic Schema
Candidate Elements = {movies, country}
8
Base Schema: IMDb
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
9
! Empty Synthetic Schema
Candidate Elements = E
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
1-size Synthetic Schema
2-size Synthetic Schema
Candidate Elements = {movies, country}
Candidate Elements = {distributor2country, country2sfx, certificate, prodcompany2country, releasein, located, shotin, location, movies, country}
10
Base Schema: IMDb
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
11
! Empty Synthetic Schema
1-size Synthetic Schema
2-size Synthetic Schema
Candidate Elements = {movies, country}
Candidate Elements = {distributor2country, country2sfx, certificate, prodcompany2country, releasein, located, shotin, location, movies, country}
Candidate Elements = E
3-size Synthetic Schema
Candidate Elements = {distributor2country, country2sfx, certificate, prodcompany2country, releasein, located, shotin, location, movies, country} Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
12
Base Schema: IMDb
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
13
! Empty Synthetic Schema
1-size Synthetic Schema
Candidate Elements = E
Candidate Elements = {movies, country}
3-size Synthetic Schema
Candidate Elements = {distributor2country, country2sfx, certificate, prodcompany2country, releasein, located, shotin, location, movies, country} Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
2-size Synthetic Schema
Candidate Elements = {distributor2country, country2sfx, certificate, prodcompany2country, releasein, located, shotin, location, movies, country} 4-size Synthetic Schema
14
• Modifying operations – Removal of properties – Insertion of new properties – Replacement of element label – Replacement of property data type
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
15
" 4-size Synthetic Schema
MOVIES Castcoverage Crewcoverage Runtimes
4-size Synthetic Schema
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
16
# Movie Ontology
4-size Synthetic Schema
MOVIES
Stars Synopsis
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
Title, Year, Stars, ColorType, Synopsis, Position, Runtimes, Opinion, Duration, Trailer, MovieURL, Episode, Credits...
MOVIES ≡ FILMS MOVIES ≡ MOVINGPICTURE
17
"
$ Movie Ontology
4-size Synthetic Schema
MOVIES
MOVIES ≡ FILMS MOVIES ≡ MOVINGPICTURE
FILMS
TELEFILMS ⊆ MOVIES MOVIES ⊆ PRODUCTION
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
18
"
%
4-size Synthetic Schema
MOVIES
VOTES NUMBER
NUMBER ⊆ NUMERIC DECIMAL ⊆ NUMERIC INTEGER ⊆ NUMERIC Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
INTEGER 19
(a) Generated Schema
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
(b) Modified Schema
20
&
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
21
& Duration (6 Elements) Elapsed Time (in sec.)
Duration (4 Elements)
Elapsed Time (in sec.)
50,00 40,00 30,00 20,00 10,00 0,00 1
2
3
4
50,00 40,00 30,00 20,00 10,00 0,00 1
2
#Schema Modifiers 10 Schemas
20 Schemas
3
4
#Schema Modifiers 10 Schemas
30 Schemas
20 Schemas
30 Schemas
Elapsed Time (in sec.)
Duration (8 Elements) 50,00 40,00 30,00 20,00 10,00 0,00 1
2
3
4
#Schema Modifiers 10 Schemas
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
20 Schemas
30 Schemas
22
Global Schema Similarity
& 1,0 0,9 0,8 0,7 0,6 0,5 0,4 0,3 0,2 0,1 0,0
1
2
3
4
4-size Schemas
0,85
0,83
0,83
0,78
6-size Schemas
0,89
0,86
0,85
0,81
8-size Schemas
0,88
0,88
0,86
0,83
#Schema Modifiers 4-size Schemas
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
6-size Schemas
8-size Schemas
23
"
'
(
• No similar work on generating synthetic database schemas in the literature • Found only tools that produce random data to populate empty schemas for performance testing purposes on databases • These tools can be used to populate the synthetic schemas Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
24
• Automatic process to generate multiple synthetic database schemas • Used in a semantic peer clustering process • Ongoing research issues – Produce synthetic queries to be executed at the synthetic schemas – Generate synthetic schemas considering multiple base schemas – Develop new schema modifiers, e.g. introduce noise in the synthetic schemas Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
25
)%
(
$
Pires, Vieira, Saraiva, Barbosa @ DEXA 2011
45
