Semantics in Retrieval by Gan Keng Hoon
on 19th October 2017 for School of Computer Sciences Staff Seminar
Talk Theme
What do we do.
Anything concerning Semantics in Retrieval
Why are we doing it.
Making information access more Natural and Relevant.
Talk Outline
Research
* Aspect Mining *Semantic-Syntax Query Model
Team
* Postgraduate
Project
* Faceted Search *Sentiment Dictionary
Collaboration
Research: Connecting Two Ends
Who is the most
active researcher in
the area of security
for the past three
years?
Which hotel is the
cleanest and near
to a mall at KB?
A list of hotels and
their reviews,
photos, ratings etc.
A list of experts
and their scholarly
articles etc.
Research: Semantics in Resources
A list of hotels and
their reviews,
photos, ratings etc.
A list of experts
and their scholarly
articles etc.
Document
Paragraph
Sentence
Phrase
Word
Object
Attribute
Value
Article
Section
Attribute
Value
Research: Semantics in Resources
Research: Area in Sentiment Analysis
Cititel
[hotel]
Room
[aspect]
RM350 [price] clean
[sentiment] 0.750
[polarity]
Object
Attribute
Value
Research: Area in Sentiment Analysis
ASPECT SENTIMENT SPATIAL
POLARTIY
POLARTIY
Text Mining
Information
Extraction
Contextual
Analysis
Research: Text Mining
Discover something that you do not know.
• Natural Language Processing
• Entity Discovery
• Relation Discovery
Research: Aspect Mining
A domain independent approach for identifying explicit opinionated
features and attributes that are strongly related
Published Work: Saif Addeen Alrababah, Gan Keng Hoon, Tan Tien
Ping: Mining Opinionated Product Features using WordNet
Lexicographer Files. Journal of Information Science: SAGE (2016).
Research: Aspect Mining
Improve relevancy of mined items
Using aspect ranking by merging sentiment analysis and TOPSIS
(Technique for Order Performance by Similarity to Ideal Solution)
Research: Aspect Mining
Aspect1
Aspect2 Aspectm
Freq(A)
OS(A)
Relevancy(A)
Research: Aspect Mining
Aspect ranking using MCDM, i.e. Topsis and Vikor.
Improvisation of MCDM.
Research Article: Saif Addeen Alrababah, Gan Keng Hoon, Tan Tien
Ping: Comparative Analysis of MCDM Methods for Product Aspect Ranking:
TOPSIS and VIKOR. International Conference on Information and
Communication Systems, ICICS 2017, 4-6 April, 2017, Irbid, Jordan.
Research Article: Saif Addeen Alrababah, Gan Keng Hoon, Tan Tien
Ping: Product Aspect Ranking using Sentiment Analysis and TOPSIS. The Third
International Conference on Information Retrieval and Knowledge
Management, CAMP 2016, 23-24 August, 2016, Melaka, Malaysia. [Best
Paper]
Research: Revisit Area in Sentiment Analysis
ASPECT SENTIMENT SPATIAL
POLARTIY
POLARTIY
Text Mining
Information
Extraction
Contextual
Analysis
Research: Big Picture
Aspect/Opinion
Mining
Aspect-Sentiment
Extraction
Polarity Scoring
Text
Mining
Information
Extraction
Contextual
Analysis
Text
Classification
Aspect/Review
Classification
Knowledge
Representation
Semantics in Resources Semantics in Query
Sentiment Analysis Scientific Articles
Search
Facet-Value
Extraction
Article
Classification
Semantic Relation
Extraction
Research Outcome
Examples
MCDM-based
Aspect Ranking
Framework
Sem-Syn Query
Model
A-S Extraction Rule
Content
Annotation
Domain A-S Lex
Unstructured to
Structured
Natural Language
Query Interface
Structured Query
Construction
Unstructured-
Structured Data
Integration
Research: Semantics in Query
Interpreting keywords query
Construct structured query
Research Article: Gan Keng Hoon, Phang Keat Keong: A query transformation
framework for automated structured query construction in structured retrieval
environment. Journal of Information Science 40(2): 249-263, SAGE (2014).
Research Article: Gan Keng Hoon, Phang Keat Keong: Finding Target and Constraint
Concepts for XML Query Construction. International Journal of Web Information
Systems 11(4): Emerald Insight (2015)
Research: Semantics in Query
Flexibility to constructing
different types of structured query
Figure showing the semantic
representation
that can be constructed into NEXI
query.
Research Article: Gan Keng Hoon,
Phang Keat Keong: A Semantic-Syntax
Model for XML Query Construction.
International Journal of Web
Information Systems 13(2): 155-172,
Emerald Insight (2017).
Team: Postgraduate
Text
Mining
Information
Extraction
Contextual
Analysis
Text
Classification
Knowledge
Representation
Semantics in Resources Semantics in Query
Sentiment Analysis
Content
Annotation
TK (Msc Mix)
Identification of Supporting Sentence
for Comparative Opinion Mining
Rizvana (Msc Mix)
Rule-based Aspect-
Sentiment Pair Extraction
Saif (Phd)
Aspect-based Sentiment Analysis
using MCDM
Aspect/Opinion
Mining
Aspect-Sentiment
Extraction
Polarity Scoring
Aspect/Review
Classification
Krol (Msc Mix)
Polarity Detection for
Contrastive/Conditional
Sentence
Erum (Phd)
Spatial and Sentiment
Extraction for POI Graph
Issa (Phd)
Event-based Short Text
Classification
Team: Big Picture of Postgraduate
Aspect/Opinion
Mining
Aspect-Sentiment
Extraction
Polarity Scoring
Text
Mining
Information
Extraction
Contextual
Analysis
Text
Classification
Aspect/Review
Classification
Knowledge
Representation
Semantics in Resources Semantics in Query
Sentiment Analysis Scientific Articles
Search
Facet-Value
Extraction
Article
Classification
Semantic Relation
Extraction
Research Outcome
Examples
MCDM-based
Aspect Ranking
Framework
Intermediate
Query Model
A-S Extraction Rule
Content
Annotation
Domain A-S Lex
Unstructured to
Structured
Natural Language
Query Interface
Structured Query
Construction
Unstructured-
Structured Data
Integration
Project: Linking Research and Application
Expert Search
http://ir.cs.usm.my/exsearch3/
SummaRev: Reviews Analysis and Summarization
http://ir.cs.usm.my/summarev4hotel/admin/sentiment_dictionary.php
Collaboration/Discussion
Making research industry relevant
Research application prototyping (FYP != programmer?)
My direction vs your direction
Collaboration setting
THANK YOU
Visit our work at
ir.cs.usm.my

Semantics in Retrieval

  • 1.
    Semantics in Retrievalby Gan Keng Hoon on 19th October 2017 for School of Computer Sciences Staff Seminar
  • 2.
    Talk Theme What dowe do. Anything concerning Semantics in Retrieval Why are we doing it. Making information access more Natural and Relevant.
  • 3.
    Talk Outline Research * AspectMining *Semantic-Syntax Query Model Team * Postgraduate Project * Faceted Search *Sentiment Dictionary Collaboration
  • 4.
    Research: Connecting TwoEnds Who is the most active researcher in the area of security for the past three years? Which hotel is the cleanest and near to a mall at KB? A list of hotels and their reviews, photos, ratings etc. A list of experts and their scholarly articles etc.
  • 5.
    Research: Semantics inResources A list of hotels and their reviews, photos, ratings etc. A list of experts and their scholarly articles etc. Document Paragraph Sentence Phrase Word Object Attribute Value Article Section Attribute Value
  • 6.
  • 7.
    Research: Area inSentiment Analysis Cititel [hotel] Room [aspect] RM350 [price] clean [sentiment] 0.750 [polarity] Object Attribute Value
  • 8.
    Research: Area inSentiment Analysis ASPECT SENTIMENT SPATIAL POLARTIY POLARTIY Text Mining Information Extraction Contextual Analysis
  • 9.
    Research: Text Mining Discoversomething that you do not know. • Natural Language Processing • Entity Discovery • Relation Discovery
  • 10.
    Research: Aspect Mining Adomain independent approach for identifying explicit opinionated features and attributes that are strongly related Published Work: Saif Addeen Alrababah, Gan Keng Hoon, Tan Tien Ping: Mining Opinionated Product Features using WordNet Lexicographer Files. Journal of Information Science: SAGE (2016).
  • 11.
    Research: Aspect Mining Improverelevancy of mined items Using aspect ranking by merging sentiment analysis and TOPSIS (Technique for Order Performance by Similarity to Ideal Solution)
  • 12.
    Research: Aspect Mining Aspect1 Aspect2Aspectm Freq(A) OS(A) Relevancy(A)
  • 13.
    Research: Aspect Mining Aspectranking using MCDM, i.e. Topsis and Vikor. Improvisation of MCDM. Research Article: Saif Addeen Alrababah, Gan Keng Hoon, Tan Tien Ping: Comparative Analysis of MCDM Methods for Product Aspect Ranking: TOPSIS and VIKOR. International Conference on Information and Communication Systems, ICICS 2017, 4-6 April, 2017, Irbid, Jordan. Research Article: Saif Addeen Alrababah, Gan Keng Hoon, Tan Tien Ping: Product Aspect Ranking using Sentiment Analysis and TOPSIS. The Third International Conference on Information Retrieval and Knowledge Management, CAMP 2016, 23-24 August, 2016, Melaka, Malaysia. [Best Paper]
  • 14.
    Research: Revisit Areain Sentiment Analysis ASPECT SENTIMENT SPATIAL POLARTIY POLARTIY Text Mining Information Extraction Contextual Analysis
  • 15.
    Research: Big Picture Aspect/Opinion Mining Aspect-Sentiment Extraction PolarityScoring Text Mining Information Extraction Contextual Analysis Text Classification Aspect/Review Classification Knowledge Representation Semantics in Resources Semantics in Query Sentiment Analysis Scientific Articles Search Facet-Value Extraction Article Classification Semantic Relation Extraction Research Outcome Examples MCDM-based Aspect Ranking Framework Sem-Syn Query Model A-S Extraction Rule Content Annotation Domain A-S Lex Unstructured to Structured Natural Language Query Interface Structured Query Construction Unstructured- Structured Data Integration
  • 16.
    Research: Semantics inQuery Interpreting keywords query Construct structured query Research Article: Gan Keng Hoon, Phang Keat Keong: A query transformation framework for automated structured query construction in structured retrieval environment. Journal of Information Science 40(2): 249-263, SAGE (2014). Research Article: Gan Keng Hoon, Phang Keat Keong: Finding Target and Constraint Concepts for XML Query Construction. International Journal of Web Information Systems 11(4): Emerald Insight (2015)
  • 17.
    Research: Semantics inQuery Flexibility to constructing different types of structured query Figure showing the semantic representation that can be constructed into NEXI query. Research Article: Gan Keng Hoon, Phang Keat Keong: A Semantic-Syntax Model for XML Query Construction. International Journal of Web Information Systems 13(2): 155-172, Emerald Insight (2017).
  • 19.
    Team: Postgraduate Text Mining Information Extraction Contextual Analysis Text Classification Knowledge Representation Semantics inResources Semantics in Query Sentiment Analysis Content Annotation TK (Msc Mix) Identification of Supporting Sentence for Comparative Opinion Mining Rizvana (Msc Mix) Rule-based Aspect- Sentiment Pair Extraction Saif (Phd) Aspect-based Sentiment Analysis using MCDM Aspect/Opinion Mining Aspect-Sentiment Extraction Polarity Scoring Aspect/Review Classification Krol (Msc Mix) Polarity Detection for Contrastive/Conditional Sentence Erum (Phd) Spatial and Sentiment Extraction for POI Graph Issa (Phd) Event-based Short Text Classification
  • 20.
    Team: Big Pictureof Postgraduate Aspect/Opinion Mining Aspect-Sentiment Extraction Polarity Scoring Text Mining Information Extraction Contextual Analysis Text Classification Aspect/Review Classification Knowledge Representation Semantics in Resources Semantics in Query Sentiment Analysis Scientific Articles Search Facet-Value Extraction Article Classification Semantic Relation Extraction Research Outcome Examples MCDM-based Aspect Ranking Framework Intermediate Query Model A-S Extraction Rule Content Annotation Domain A-S Lex Unstructured to Structured Natural Language Query Interface Structured Query Construction Unstructured- Structured Data Integration
  • 21.
    Project: Linking Researchand Application Expert Search http://ir.cs.usm.my/exsearch3/ SummaRev: Reviews Analysis and Summarization http://ir.cs.usm.my/summarev4hotel/admin/sentiment_dictionary.php
  • 22.
    Collaboration/Discussion Making research industryrelevant Research application prototyping (FYP != programmer?) My direction vs your direction Collaboration setting THANK YOU Visit our work at ir.cs.usm.my