How to Jump-Start
Taxonomy Content Creation
September 24, 2015
Today’s Speakers
Mark Leher
COO
WAND, Inc.
mleher@wandinc.com
www.wandinc.com
Bryan Bell
Executive Vice President
Expert System USA, Inc.
bbell@expertsystem.com
www.expertsystem.com
IPOTESI INGOMBRO EVENTUALE
SCREENSHOT
IPOTESI INGOMBRO EVENTUALE
SCREENSHOT
Agenda:
• Define Terms: Taxonomy and Tagging
• Common Content Challenges
• Value of Metadata
• How to get started in your organization
•The value of semantic disambiguation
• Discuss and demonstrate the combined
WAND / Expert System solution
WAND history
• Building Taxonomies since 1983
• Taxonomy Library covering most industries and business topics:
Products and Services, Manufacturing, Skills, Insurance, Food Science, Banking,
Finance, Medical, Consumer Sentiment, General Business, Records Retention,
News, Legal, and more.
• Taxonomy Professional Services
• Client base includes online yellow pages, online advertising engines,
enterprise search, e-commerce, and SharePoint customers - SMBs to
Fortune 500
Expert System history
• Founded: 1989
• Global Headquarters: Modena, Italy
• US Headquarters: Chicago, IL
• Global, public traded company:
Listed on the AIM Stock exchange
• Scalable solution used by millions of
end users to increase the FINDABILITY
of strategic communications.
• 9 out of 10 of customers who purchase
Expert System software extend their
license to new business use cases
within 6 months.
• Advanced solution designed to
support multiple use cases.
Today’s primary focus: Taxonomy creation,
taxonomy enrichment and the deployment
to automated categorization solutions.
Proven, scalable software solution
Technology company: we develop, deploy and support out-of-the-box,
software solutions for a wide variety of knowledge management
applications that are flexible and apply to multiple use cases.
Deep linguistic analysis engine combined with a robust semantic network:
The largest, fastest growing company in the space.
Example taxonomy & extraction deployments:
Expert System history
IPOTESI INGOMBRO EVENTUALE
SCREENSHOT
IPOTESI INGOMBRO EVENTUALE
SCREENSHOT
What is a Taxonomy?
Collection of concepts for a
given subject domain organized
into a hierarchical tree structure
• Concepts = Categories = Terms =
Preferred Terms
• Synonyms = Non-Preferred Terms
Individual categories from the
taxonomy are tagged to
documents as metadata
IPOTESI INGOMBRO EVENTUALE
SCREENSHOT
IPOTESI INGOMBRO EVENTUALE
SCREENSHOT
Basic Elements of a Taxonomy
• Broader Term: Asset Types
• Preferred Term: Foreign
Exchange
• Synonym: Foreign Currencies
• Narrower Term: Australian
Dollar
IPOTESI INGOMBRO EVENTUALE
SCREENSHOT
IPOTESI INGOMBRO EVENTUALE
SCREENSHOT
What is Tagging?
• Tagging means applying metadata to documents
• Descriptive Metadata: One or multiple words about the
content of the document
• Tags will be search refinements for end users
IPOTESI INGOMBRO EVENTUALE
SCREENSHOT
IPOTESI INGOMBRO EVENTUALE
SCREENSHOT
Common Challenges:
• Enterprise content is exploding and findability is a major
challenge
• Folder based organization doesn’t really work. Only
metadata is scalable
• Two key ingredients that most companies are missing:
• Well defined corporate taxonomy
• Capability to tag content with taxonomy based
metadata
IPOTESI INGOMBRO EVENTUALE
SCREENSHOT
IPOTESI INGOMBRO EVENTUALE
SCREENSHOT
Make the Case:
• Companies that have taxonomies are 250% more likely to
have users that are satisfied or very satisfied with search.*
• Information organization (information architecture,
taxonomies and tagging) is the second highest priority for
investment in 2013 and 2014.*
• In an independent survey, 78% of respondents believe
finding the right information is critical or imperative to the
organization’s overall success and business goals.*
*Enterprise Search and Findability Report, Findwise, 2013/2014
IPOTESI INGOMBRO EVENTUALE
SCREENSHOT
IPOTESI INGOMBRO EVENTUALE
SCREENSHOT
Building your taxonomies:
• Understand your use case and need
• Business users need to be involved
• Talk to end users
– How do they want to search for information
– How do they think of the organization
– What terms do they use
• Which taxonomies are important for which documents?
• Don’t try to build one giant taxonomy
• Develop a tagging strategy
IPOTESI INGOMBRO EVENTUALE
SCREENSHOT
IPOTESI INGOMBRO EVENTUALE
SCREENSHOT
Ongoing Governance:
• Include Taxonomy Management in your governance
plan
• Policies and standards for when to create a new term,
definitions, managing re-used terms, and term naming
conventions.
• Change management planning
Taxonomy Studio
Content visualization
Cogito Categorizer
Automated categorization
• WAND taxonomy import
• Data driven taxonomy enrichment
• Content visualization
• Training set identification
• Automated rule generation
• Rule visualization & modification
• Content driven taxonomy deployment.
• Combining multiple data sources.
• Combining disparate data sources.
• Common metadata model.
Technology alone struggles to address
language ambiguities or understand
context…
Same
Word,
Different
Meanings
Different
Words,
Same
Meaning
Different
Words,
Related
Meaning
Press: push or the
news media?
Buy or purchase? Hollywood or the
U.S. film industry?
Content driven taxonomies using
the power of language.
Our semantic network
is a rich map of definitions
of words and associations
between words.
2 million concepts and their varied meanings
10 million relationships between these meanings
Semantic network (language ontology) is designed
to remove word &
word meaning ambiguity.
Content driven taxonomies using
the power of language.
What defines a semantic platform?
Morphological
analysis word forms dog, dog-catcher, doggy bag
Grammatical analysis parts of speech "There are 40 rows in the table." (noun)
"She rows 5 times a week." (verb)
Logical analysis
word
relationships
"The car I bought, to replace my Chrysler,
stinks."
Semantic analysis word context "I used chicken broth for my soup stock."
"I have 10,000 apples in stock."
"I bought 10,000 shares of stock in Apple."
Deep linguistic analysis of words to understand context.
Cogito Intelligence API
Dynamic metadata management, categorization, text
mining, fact mining, etc…
Dynamic metadata assignment
content driven taxonomy development
Expert System – The Disambiguator
Word disambiguation: back-end demonstration
Content input panel
Semantic and linguistic
analysis panel
Automated categorization
S-A-O representation
Not an end-user GUI
• Taxonomy import and export enabled
• Content driven taxonomies through content
visualization
- dynamic concept identification
- content driven taxonomy node suggestions
- training set identification
• Automatic rule creation
• Rules can be exported to Cogito Studio for viewing and
modified if required.
• Taxonomies are deployed into Cogito Categorizer for
automated categorization.
Future:
Taxonomy Studio and Cogito Studio are being integrated as
a combined solution.
- outcome will be a Taxonomy / Ontology Suite
Expert System – Taxonomy Studio
Taxonomy Studio
Content visualization
WAND Human Resources taxonomy
Expert System tag cloud
- data visualization-
Main Lemmas:
Dynamic visualization
Key word / Concept search
Taxonomy Studio
Content visualization for content driven taxonomy support
Main lemma tag cloud
Taxonomy Studio
content visualization
Taxonomy Studio
OWL, SKOS and .TXT export standards compliant
Thank you
Mark Leher
mleher@wandinc.com
www.wandinc.com
Bryan Bell
bbell@expertsystem.com
www.expertsystem.com

How to Jump Start Taxonomy Content Creation webinar slides 9 24 15

  • 1.
    How to Jump-Start TaxonomyContent Creation September 24, 2015
  • 2.
    Today’s Speakers Mark Leher COO WAND,Inc. mleher@wandinc.com www.wandinc.com Bryan Bell Executive Vice President Expert System USA, Inc. bbell@expertsystem.com www.expertsystem.com
  • 3.
    IPOTESI INGOMBRO EVENTUALE SCREENSHOT IPOTESIINGOMBRO EVENTUALE SCREENSHOT Agenda: • Define Terms: Taxonomy and Tagging • Common Content Challenges • Value of Metadata • How to get started in your organization •The value of semantic disambiguation • Discuss and demonstrate the combined WAND / Expert System solution
  • 4.
    WAND history • BuildingTaxonomies since 1983 • Taxonomy Library covering most industries and business topics: Products and Services, Manufacturing, Skills, Insurance, Food Science, Banking, Finance, Medical, Consumer Sentiment, General Business, Records Retention, News, Legal, and more. • Taxonomy Professional Services • Client base includes online yellow pages, online advertising engines, enterprise search, e-commerce, and SharePoint customers - SMBs to Fortune 500
  • 5.
    Expert System history •Founded: 1989 • Global Headquarters: Modena, Italy • US Headquarters: Chicago, IL • Global, public traded company: Listed on the AIM Stock exchange • Scalable solution used by millions of end users to increase the FINDABILITY of strategic communications. • 9 out of 10 of customers who purchase Expert System software extend their license to new business use cases within 6 months. • Advanced solution designed to support multiple use cases. Today’s primary focus: Taxonomy creation, taxonomy enrichment and the deployment to automated categorization solutions.
  • 6.
    Proven, scalable softwaresolution Technology company: we develop, deploy and support out-of-the-box, software solutions for a wide variety of knowledge management applications that are flexible and apply to multiple use cases. Deep linguistic analysis engine combined with a robust semantic network: The largest, fastest growing company in the space. Example taxonomy & extraction deployments: Expert System history
  • 7.
    IPOTESI INGOMBRO EVENTUALE SCREENSHOT IPOTESIINGOMBRO EVENTUALE SCREENSHOT What is a Taxonomy? Collection of concepts for a given subject domain organized into a hierarchical tree structure • Concepts = Categories = Terms = Preferred Terms • Synonyms = Non-Preferred Terms Individual categories from the taxonomy are tagged to documents as metadata
  • 8.
    IPOTESI INGOMBRO EVENTUALE SCREENSHOT IPOTESIINGOMBRO EVENTUALE SCREENSHOT Basic Elements of a Taxonomy • Broader Term: Asset Types • Preferred Term: Foreign Exchange • Synonym: Foreign Currencies • Narrower Term: Australian Dollar
  • 9.
    IPOTESI INGOMBRO EVENTUALE SCREENSHOT IPOTESIINGOMBRO EVENTUALE SCREENSHOT What is Tagging? • Tagging means applying metadata to documents • Descriptive Metadata: One or multiple words about the content of the document • Tags will be search refinements for end users
  • 10.
    IPOTESI INGOMBRO EVENTUALE SCREENSHOT IPOTESIINGOMBRO EVENTUALE SCREENSHOT Common Challenges: • Enterprise content is exploding and findability is a major challenge • Folder based organization doesn’t really work. Only metadata is scalable • Two key ingredients that most companies are missing: • Well defined corporate taxonomy • Capability to tag content with taxonomy based metadata
  • 11.
    IPOTESI INGOMBRO EVENTUALE SCREENSHOT IPOTESIINGOMBRO EVENTUALE SCREENSHOT Make the Case: • Companies that have taxonomies are 250% more likely to have users that are satisfied or very satisfied with search.* • Information organization (information architecture, taxonomies and tagging) is the second highest priority for investment in 2013 and 2014.* • In an independent survey, 78% of respondents believe finding the right information is critical or imperative to the organization’s overall success and business goals.* *Enterprise Search and Findability Report, Findwise, 2013/2014
  • 12.
    IPOTESI INGOMBRO EVENTUALE SCREENSHOT IPOTESIINGOMBRO EVENTUALE SCREENSHOT Building your taxonomies: • Understand your use case and need • Business users need to be involved • Talk to end users – How do they want to search for information – How do they think of the organization – What terms do they use • Which taxonomies are important for which documents? • Don’t try to build one giant taxonomy • Develop a tagging strategy
  • 13.
    IPOTESI INGOMBRO EVENTUALE SCREENSHOT IPOTESIINGOMBRO EVENTUALE SCREENSHOT Ongoing Governance: • Include Taxonomy Management in your governance plan • Policies and standards for when to create a new term, definitions, managing re-used terms, and term naming conventions. • Change management planning
  • 14.
    Taxonomy Studio Content visualization CogitoCategorizer Automated categorization • WAND taxonomy import • Data driven taxonomy enrichment • Content visualization • Training set identification • Automated rule generation • Rule visualization & modification • Content driven taxonomy deployment. • Combining multiple data sources. • Combining disparate data sources. • Common metadata model.
  • 15.
    Technology alone strugglesto address language ambiguities or understand context… Same Word, Different Meanings Different Words, Same Meaning Different Words, Related Meaning Press: push or the news media? Buy or purchase? Hollywood or the U.S. film industry? Content driven taxonomies using the power of language.
  • 16.
    Our semantic network isa rich map of definitions of words and associations between words. 2 million concepts and their varied meanings 10 million relationships between these meanings Semantic network (language ontology) is designed to remove word & word meaning ambiguity. Content driven taxonomies using the power of language.
  • 17.
    What defines asemantic platform? Morphological analysis word forms dog, dog-catcher, doggy bag Grammatical analysis parts of speech "There are 40 rows in the table." (noun) "She rows 5 times a week." (verb) Logical analysis word relationships "The car I bought, to replace my Chrysler, stinks." Semantic analysis word context "I used chicken broth for my soup stock." "I have 10,000 apples in stock." "I bought 10,000 shares of stock in Apple." Deep linguistic analysis of words to understand context.
  • 18.
    Cogito Intelligence API Dynamicmetadata management, categorization, text mining, fact mining, etc…
  • 19.
    Dynamic metadata assignment contentdriven taxonomy development
  • 20.
    Expert System –The Disambiguator Word disambiguation: back-end demonstration Content input panel Semantic and linguistic analysis panel Automated categorization S-A-O representation Not an end-user GUI
  • 21.
    • Taxonomy importand export enabled • Content driven taxonomies through content visualization - dynamic concept identification - content driven taxonomy node suggestions - training set identification • Automatic rule creation • Rules can be exported to Cogito Studio for viewing and modified if required. • Taxonomies are deployed into Cogito Categorizer for automated categorization. Future: Taxonomy Studio and Cogito Studio are being integrated as a combined solution. - outcome will be a Taxonomy / Ontology Suite Expert System – Taxonomy Studio
  • 22.
    Taxonomy Studio Content visualization WANDHuman Resources taxonomy Expert System tag cloud - data visualization- Main Lemmas: Dynamic visualization Key word / Concept search
  • 23.
    Taxonomy Studio Content visualizationfor content driven taxonomy support Main lemma tag cloud
  • 24.
  • 25.
    Taxonomy Studio OWL, SKOSand .TXT export standards compliant
  • 26.
    Thank you Mark Leher mleher@wandinc.com www.wandinc.com BryanBell bbell@expertsystem.com www.expertsystem.com