Title: A Heuristic-based Methodology for Semantic Augmentation of User Queries on the Web
1A Heuristic-based Methodology for Semantic
Augmentation ofUser Queries on the Web
- Veda C. Storey
- Georgia State University
- October 2003
2The worlds largest information resource
Over 3 billion pages
3What we know about web search
- Average query length 2 words
- Advanced syntax lt5 of queries
- Syntax often used incorrectly
- Query results
- many irrelevant
- most users look at first 2-8
- users modify query gt50 of the time
- Source Spink et al. (2001 2002)
4The need for precision
2 out of first 10 pages
5The problem of semantics
6Agenda
- Research Objectives
- Heuristic-based Methodology
- Prototype
- Testing
- Future Research
7Research Objectives
- Objectives
- develop agent-based methodology for web
information retrieval that adds context to a
users query - ontologies
- lexical resources
- implement methodology in prototype and evaluate it
8Related Research
9Semantic Web Example Query
- Mom needs to see a specialist and then has to
have a series of physical therapy sessions.
Biweekly or somethingset up the appointments. - What knowledge does agent need?
- Meaning of terms specialist, series, physical
therapy, session, bi-weekly, appointments - Constraints Moms medical history, distance from
doctor, weekly schedule, insurance plan
10Example
- Mom needs to see a specialist and then has to
have a series of physical therapy sessions.
Biweekly or somethingset up the appointments. - Dr. Mullen
- 2476 Northridge Ave., Atlanta
- 404-776-2990
- Dr. Pillbury
- 7227 Dunwoody Club Dr. Suite 201, Atlanta
- 770-455-2244
- Dr. Kenton
- 81 Forest Cres., Atlanta
- 678-277-4981
11Methodology Architecture
12WordNet versus DARPA Ontologies
DARPA Ontology Library Six ontologies found with
class Professor. Extract from
www.daml.org/ontologies/64 Faculty (teacherOf).
Professor (age, doctoralDegreeFrom,
emailAddress, mastersDegreeFrom,
researchInterest, undergraduateDegreeFrom,
tenured) AssistantProfessor () AssociateProfessor
() Chair () Dean () FullProfessor
() VisitingProfessor ()
WordNet 1.7.1 Synonyms (professor means the
same as x) null Hypernyms (professor is a kind
of x) academician, academic, faculty member Hypo
nyms (x is a kind of professor) assistant
professor, associate professor, full professor,
Regius professor, visiting professor Holonyms
(professor is a part of x) member of faculty,
member of staff Familiarity (professor has x
word senses) one
13WordNet Search on Chair
- The noun "chair" has 4 senses in WordNet.1.
chair -- (a seat for one person, with a support
for the back "he put his coat over the back of
the chair and sat down")2. professorship, chair
-- (the position of professor "he was awarded an
endowed chair in economics")3. president,
chairman, chairwoman, chair, chairperson -- (the
officer who presides at the meetings of an
organization "address your remarks to the
chairperson")4. electric chair, chair, death
chair, hot seat -- (an instrument of execution by
electrocution resembles a chair "the murderer
was sentenced to die in the chair") - The verb "chair" has 2 senses in WordNet.1.
chair, chairman -- (act or preside as chair, as
of an academic department in a university "She
chaired the department for many years")2.
moderate, chair, lead -- (preside over "John
moderated the discussion")Search for of senses
Show glossesShow contextual help
14DAML Ontologies
Term Hypernym Ontological Information Source
Specialist Doctor Specialist is a subclass of Medical Practice Doctor is a subclass of Medical care http//www.daml.org/ontologies/106 http//www.daml.org/ontologies/132
Physical therapy Therapy Physical therapy services is a subclass of Medical practice Drug therapy is a subclass of Medical Treatment Event Medical treatment event is a subclass of Medical care http//www.daml.org/ontologies/106 http//www.daml.org/ontologies/179
Series Ordering World Series game is a subclass of Event Ordering is a subclass of Social Interaction http//www.daml.org/ontologies/195 http//www.daml.org/ontologies/172
Appointment Meeting Session Meeting is a subclass of Social Interaction Meeting is a subclass of Event http//www.daml.org/ontologies/172 http//www.daml.org/ontologies/4
15Methodology Evolving Semantic
Network
- Semantic network for a two-word query
he
c
s
n
Key he hypernym n negation c candidate
s synonym
16Methodology Evolving Semantic
Network
- Example Queries
- Find a chair of a university
- Find a chair in Atlanta
- Find doctors who provide physical therapy
17Knowledge base Build semantic network for query
terms
- Semantic network created
- Heuristic 3 Create candidate link between
sequential terms in query
Query 1
Query 2
Query 3
18Inferencing Expand knowledge of query terms
- Heuristic 4 Phrases
- WordNet queried for each ordered pair
- Chair University no
- Chair Atlanta no
- Doctor Physical no
- Physical Therapy yes
19Inferencing Expand knowledge of query terms
- Heuristic 5 Knowledge access for chair
20Inferencing Expand knowledge of query terms
- Heuristic 6 Match terms in knowledge base
Requires user to identify context
he
n
Key he hypernym n negation c candidate
21Heuristics Construct Query
- Example context aware query
- (Chair and Professor) (University or College) not
Seat
he
s
n
Key he hypernym n negation c candidate
s synonym
22ISRA (Intelligent Semantic Web Retrieval Agent)
Server Side
Client Side
Web Services
Query Refinement And Execution Agent
23Prototype
- Interface agent Accept and parse query
24Prototype
- Knowledge Inferencing agent Synonyms Negation
from WordNet
25Prototype
- Knowledge Inferencing agent Hypernyms from
WordNet DARPA Ontology Library
26(No Transcript)
27 Sample Query
Search Engine ISRA Search Engine
Query Import review Query IMPORT andnot meaning and goods and REVIEW andnot reappraisal and proceeding
Hits 2,402,059 Hits 31,041
Relevant Pages 20 Not Relevant Relevant Pages 15 Relevant, 5 Partially Relevant
Dreamcast.ign.com Giga Wing (Import) Review Continue to Giga Wing (Import) this session brought to you by about this ad continue to Giga Wing (ImportDescriptionIncludes review and screenshots.http//dreamcast.ign.com/articles/162/162809p1.html Import Export Companies on Internet Import Export Companies on Internet ... modified on 05/23/03 200504 Import/Export Business You are welcomed ... spices.. We are AFERIAD Co. Import Export, specialised in the ... milkpowder, juice, spices and Goods. We also export canned fish ... http//shinestar.com/newera/links.htm
28Results
- ISRA helps Using ISRA in conjunction with
search engine significantly increases relevance - Search engines matter Google returns
significantly more relevant pages than AlltheWeb - Ambiguity matters Queries with more precise
terms return significantly more relevant pages - Interactions matter
- ISRA provides greater increase in relevance when
used with Google than AlltheWeb - ISRA provides greater increase in relevance when
search terms are ambiguous
29Methodology Assumptions
- 1. Not require excessive user intervention
- 2. Knowledge sources trustworthy
- Expanded methodology
30Conclusion
- Semantic Web
- Methodology and tools to advance Semantic Web
- Heuristic-based methodology
- Preliminary results supportive
- Future
- Prototype
- User profiles
- Testing