Semantic Web Techniques for Personalization of eGovernment Services

1
Semantic Web Techniquesfor Personalizationof
eGovernment Services
SemWAT 2006 1st International ER Workshop on
Semantic Web Applications Theory and Practice
Tucson, AZ, November 2006
Federica Mandreoli Riccardo Martoglia Enrico
Ronchetti Paolo Tiberio Università degli Studi
di Modena e Reggio Emilia
Fabio Grandi Maria Rita Scalas Università degli
Studi di Bologna
2
Overview

• Our research activities concern the
  implementation of Web information systems for
  eGovernment applications
• Due to development of eGovernment
  initiatives,more and more on-line resources and
  services are being made available by Public
  Administrations (PAs)
• We make use of temporal database and semantic Web
  techniques to provide personalized access to such
  resources and services
• In particular, we consider multi-version norm
  texts (stored in XML format) available in Web
  repositories

3
Importance of versioning

• Temporal concerns are ubiquitous in the law
  domain
• Each normative text changes in time due to
  different modifications, but keeps its identity
• The ability to model temporal dimensions is
  essential for the management of evolving norms
• it is crucial to reconstruct the consolidated
  version of a norm
• also past versions are still important

4
Importance of versioning

• Applicability (semantic) versioning also plays an
  important role
• some norms or some of their parts have or acquire
  a limited applicability
• personalized version of the norm
• A version only containing provisions which are
  applicable to a citizens personal case

Art. 1 (unemployed) xxy yyx yxyx yyyxx
xyyx Art. 2 (self-employed) aab bbab abab
abba ab Art. 3 (retired) qwqq ww wqqw wq ww
Self-employed
5
Motivation

• Large XML collections of norms are made
  available by the PA on the Web but
  personalization is
• Absent, e.g. http//www.normeinrete.it(temporal
  versioning partially supported)
• Predefined in the Website structure and contents,
  e.g. http//www.italia.gov.it(hardwired by
  human experts following the life-events
  approach)
• Lack of an effective, flexible, on-demand
  (intelligent, efficient) personalization
  facility

6
Objectives

• Development of an effective and efficient Web
  information system where
• norms are represented as XML documents
• dynamics of norms in time is captured
• limited applicability of norms (and their parts)
  is captured
• selective access and reconstruction of
  versionsis supported by a query engine
• Aimed at
• enabling citizens to access personalized versions
  of multiversion resources
• improving and optimizing the involvement of
  citizens in the eGovernance process

7
The Technological Infrastructure
Public Administration DB
WEB SERVICES OF PUBLIC ADMINISTRATION
1
WEB SERVICES WITH ONTOLOGY OC
2
SIMPLE ELABORATION UNIT
creation /update
class Cx
1 identification phase reconstruction
on-the-fly of the digital identity of the
authenticated user
2 classification phase use of the collected
digital identity to classify the citizen with
respect to the civic ontology Oc
3 querying phase access and reconstruction of
all and only norms which are applicable to the
class Cx
8
The Civic Ontology

• Embodies a classification of citizens based on
  the distinctions introduced by successive norms
  that imply some limitations in their
  applicability (founding acts)
• At this stage of the project, we manage
  tree-like ontologies(i.e. class taxonomies
  induced by the IS-A relationship)

9
The modeling approach

• Extension of a previous temporal XML model (DKE
  2005) including
• a temporal multi-version XML schema
• is based on the hierarchical organization of
  normative texts contents-section-article-paragrap
  h
• at each level of the hierarchy, the history of
  changes is represented by the (time-stamped)
  versions produced
• it supports ancestor-descendant inheritance
• temporal manipulation operations
• Addition of applicability annotations in order to
  support semantic versioning

10
The temporal XML schema
Num R
Law

• 4 Temporal Dimensions
• Publication time
• time of publication on the Official Journal
• Validity time
• time the norm is in force
• Efficacy time
• time the norm can be applied
• Transaction time
• time the norm is storedin the system

Type R
Title
Contents
An_ref O
Ver
Num R
Section
Num R
Ver
An_ref O
Num R
Heading
Num R
Article
Ver
An_ref O
Num R
Paragraph
Heading
Num R
Ver
An_ref O
Num R
11
Semantic versioning

• A pre-order and post-order numbering scheme is
  introduced in the tree-like ontology
• Classes are identified by means of their
  pre-order code
• Encoding is exploited in query processing for
  quick ancestor-descendant checking
• Applicability annotations (AA) are added to
  semantic versions of document parts as references
  to the ontology classes

12
Semantic versioning

• Applicability is inherited by descendant nodes
  unless locally redefined
• By means of redefinitions we can also introduce,
  for each part of a document, complex
  applicability properties
• Restrictions with respect to ancestors
• Extensions with respect to ancestors

13
Example of full search

• John Smith is a self-employed citizen.
• He is interested in the text of all the norms ...
• ... which contain paragraphs dealing with health
  care, ...
• ... which were valid and in effect between 2002
  and 2004, ...
• ... and which are applicable to his case (civic
  class 7).

Structural constraint Textual constraint Temporal
constraint Semantic constraint
4 orthogonal constraints
14
Example of full search
FOR a IN norms WHERE textConstr
(a//paragraph//text(), health AND care) AND
tempConstr (vTime OVERLAPS PERIOD(2002-01-01,2
004-12-31)) AND tempConstr (eTime OVERLAPS
PERIOD(2002-01-01,2004-12-31)) AND
applConstr (class 7) RETURN a
Structural constraint Textual constraint Temporal
constraint Semantic constraint
4 orthogonal constraints
15
Example of full search
Civic ontology
Normative DB
Norm
Article 1
Article 2
TA
Ver 1
AA3
Par 1
Par 2
norm//paragraph//text()
TA
TA
TA
Ver 1
Ver 1
Ver 2
AA4
AA3,8
class 7
AA
Health care text X
Public health text Y
Health care text Z
16
Our prototype system (native approach)

• The query engine is able to access and retrieve
  only the strictly necessary data
• selection relies on ad-hoc data structures
  supporting multi-versioning
• storage granularity is finer than the entire
  documents used by standard XML engines (including
  our previous prototype stratum approach)
• Only the parts which satisfy the temporal and
  applicability constraints are used for the
  reconstruction of the retrieved documents
• There is no need to retrieve whole XML documents
  and build space-consuming structures such as DOM
  trees

17
Evaluation benchmark

• Variable document size
• min 2KB
• avg 24KB
• max 125KB

• Three XML document sets
• 5000 documents (120MB)
• 10000 documents (240MB)
• 20000 documents (480MB)

• Five different query types
• Queries on keywords (structural textual
  constraints)
• Q1 keywords in contents
• Q2 keywords in type and contents
• Temporal queries (structural temporal
  constraints)
• Q3 conditions on publication, validity and
  transaction time
• Mixed queries (structural textual temporal
  constraints)
• Q4, Q5 with keywords and temporal conditions

• Five variants with semantic constraints
• Qx-A with additional applicability constraints

18
Performance evaluation

• The new system outperforms its predecessor
  (stratum approach) as far as temporal queries
  are concerned
• The new system showed a very high efficiencyin
  personalization query processing
• selection of qualifying versions is improved by a
  technique involving simple comparisons involving
  pre-post encodings
• 0.5-1 more time than for the original versions
• 3-4 storage space overhead
• The new system showed good scalability figures in
  every type of query context
• the computing time grows sublinearly with the
  number of documents (it depends mainly on the
  size of the results)

19
Conclusions

• We presented our research work concerning the
  design and implementation of efficient Web-based
  information systems for eGovernment applications
• We introduced support for a personalized access
  to resources on the basis of the digital
  identity of citizens (relying on semantic
  versioning and ontology mapping)
• We developed an efficient platform (native
  approach) for which a specialized Multi-version
  XML Query Processor has been designed and
  implemented
• We showed our approach to be very efficient in a
  large set of experimental situations with good
  scale-up figures under growing load configurations

20
Future Work

• Extensions of the current framework
• more advanced application requirements may
  include a more sophisticated ontology definition
  (graph-like), possibly versioned, and more
  advanced reasoning services
• Completion of the technological infrastructure
  usable in a large Web-based eGovernment scenario,
  including
• identification and classification services
• Assessment of our prototype systems in a concrete
  working environment
• with real users and with a large repository of
  real norms
• Extension to a more general application
  domain(Web personalization via ontology-based
  user profiling)