P17-A Methodology for Developing a Taxonomy for an ...


Published on

Published in: Education, Technology
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide
  • Taxonomy in its broadest sense involves classifying things such as the original effort to classify living things (Carl Linnaeus in the 1700s) More recently, taxonomy is being used by organizations to classify things that are important to an organization, corporation or government agency.
  • Semantic web – From Wikipedia - The Semantic Web is a project aimed to make web pages understandable by computers, so that they can search websites and perform actions in a standardized way. It emphasizes information exchange by giving meaning ( semantics ), in a manner understandable by machines, to the content of documents on the Web. The Semantic Web extends the World Wide Web through the use of standards, markup languages and related processing tools.
  • Conceptual data model vs. Logical data model – varies by methodology; different levels of detail; conceptual may have many to many relationships and limited or no attributes Physical data model – design of a physical database, usually in data definition language using SQL commands.
  • P17-A Methodology for Developing a Taxonomy for an ...

    1. 1. A Methodology for Developing a Taxonomy – A Subject Oriented Approach International Symposium on Ontology-Metamodeling State Key Laboratory of Software Engineering, Wuhan University Richard Jordan, Computer Specialist, Office of the Chief Information Officer, Federal Aviation Administration United States of America (USA) With contributions from Kirk Lutz, IBM Corporation March 2006 Presented to: By: Date:
    2. 2. Objectives/Outline <ul><li>Our Context: Introduction & Drivers </li></ul><ul><li>Deriving an FAA Taxonomy </li></ul><ul><li>Functional Taxonomies </li></ul>
    3. 3. 1. Taxonomy for Organizations: Introduction & Drivers <ul><li>Providing a classification of information stored in many different forms – relational data, documents, digital assets, XML, web pages, web services, discussion groups, etc. </li></ul><ul><ul><li>By tagging such assets with relevant terms from the taxonomy, we enable search and retrieval of those information assets </li></ul></ul><ul><ul><li>Getting users to the content they need – quickly </li></ul></ul><ul><li>Taxonomies are: </li></ul><ul><ul><li>often hierarchical, sometimes a network structure </li></ul></ul><ul><ul><li>Used often for web content management </li></ul></ul><ul><ul><li>Considered important for having “semantic” web capabilities </li></ul></ul>
    4. 4. Strategic Drivers and Context for Taxonomies in Government <ul><li>E-Government – making government more accessible to citizens through the Internet and automated capabilities </li></ul><ul><li>Enterprise Content and Data management – </li></ul><ul><ul><li>Growing needs for these capabilities including metadata management and effective access to data & web resources </li></ul></ul><ul><ul><li>Often viewed as separate disciplines </li></ul></ul><ul><ul><li>Data Sharing is a driver – part of data management </li></ul></ul><ul><li>Making the Internet a better resource – “The Semantic Web” </li></ul>
    5. 5. USA - Federal Data Reference Model (DRM) <ul><li>One of the Reference Models making up the framework for the Federal Enterprise Architecture (FEA) </li></ul><ul><li>DRM Version 2 – three parts: </li></ul><ul><ul><li>Data Description – entities, attributes, and relationships </li></ul></ul><ul><ul><li>Data Sharing - Information Exchange Packages </li></ul></ul><ul><ul><li>Data Context – Taxonomy, Ontology, Classification </li></ul></ul><ul><li>Data Context part calls for U.S. government agencies to have a method, such as a taxonomy or ontology, to enable its customers to search for and retrieve information </li></ul>
    6. 6. 2. Deriving an FAA Taxonomy <ul><li>Corporate Data Architecture: A model of the data objects that are relevant to an enterprise, their relationship to each other, and the principles and guidelines governing their design and evolution over time. </li></ul><ul><li>Scope: FAA-wide </li></ul><ul><li>A part of the FAA Enterprise Architecture </li></ul><ul><li>In Entity-Relationship format </li></ul>
    7. 7. Methodology <ul><li>Form a logical subject area centered on a kernel entity </li></ul><ul><li>A kernel entity represents a business object that stands alone and is not dependent on any other entity </li></ul><ul><ul><li>Examples: Flight Event, Person, & Course </li></ul></ul><ul><li>Each subject area is named for a kernel entity (based on information engineering methodology) </li></ul><ul><ul><li>Some subtype entities under a kernel entity are so complex, we separate them into their own sub-subject area </li></ul></ul><ul><li>The subject areas make up a logical data model </li></ul><ul><li>Collect similar subject areas into higher level subject areas where needed </li></ul><ul><ul><li>Example: Parties is a higher level subject area that encompasses Person and Organization </li></ul></ul><ul><li>Iterate top-down and bottom-up to complete the analysis </li></ul><ul><li>This represents a subject-oriented (data centric) hierarchical taxonomy of an organization. </li></ul><ul><li>Some data instances are categorized using valid values for reference entities (for example, an instance of Aircraft Type is glider, balloon, blimp/dirigible, fixed wing single engine, rotorcraft, etc.) </li></ul>
    8. 8. Portions of FAA Taxonomy <ul><li>Parties </li></ul><ul><ul><li>Organizations </li></ul></ul><ul><ul><li>Organization Positions </li></ul></ul><ul><ul><li>Persons </li></ul></ul><ul><li>Events </li></ul><ul><ul><li>Flight Events </li></ul></ul><ul><ul><li>Flight Plan Events </li></ul></ul><ul><ul><li>Weather Observations </li></ul></ul>
    9. 9. This Kind of Taxonomy <ul><li>In our subject oriented taxonomy, the terms are: </li></ul><ul><ul><li>Complete – to the extent that logical data modeling is complete, then the taxonomy is complete </li></ul></ul><ul><ul><li>Either non-redundant or are subtype of higher terms </li></ul></ul><ul><ul><li>Consistent kinds of terms – all nouns – may facilitate end user usage and hit rate </li></ul></ul><ul><li>This kind of taxonomy: </li></ul><ul><ul><li>Not especially designed for web search, retrieval or navigation </li></ul></ul><ul><ul><li>Provides completeness </li></ul></ul><ul><ul><li>Enables metadata management including data classification </li></ul></ul><ul><li>Aliases can augment this kind of taxonomy </li></ul>
    10. 10. 3. Functionally Oriented Taxonomies <ul><li>Many taxonomies for accessing web resources use functional terms (rather than nouns or entities) to approximate the purpose or need of end users </li></ul><ul><li>These are often service oriented and “citizen-centric” – examples </li></ul><ul><ul><li>Finding a national park with swimming </li></ul></ul><ul><ul><li>Applying for a pilot’s license </li></ul></ul><ul><ul><li>Finding known pollution sites near my address </li></ul></ul><ul><li>U.S. government is calling for use of the process part of its Federal Enterprise Architecture (called the Business Reference Model or BRM) to be used by federal agencies as a taxonomy </li></ul>
    11. 11. USA’s Business Reference Model (BRM) <ul><li>Organized into 3 tiers </li></ul><ul><ul><li>Business Area </li></ul></ul><ul><ul><ul><li>Line of Business </li></ul></ul></ul><ul><ul><ul><ul><li>Sub-function </li></ul></ul></ul></ul><ul><li>Example: </li></ul><ul><ul><li>Business Area: Transportation </li></ul></ul><ul><ul><ul><li>Line of Business: Air Transportation </li></ul></ul></ul><ul><ul><ul><ul><li>Sub-function: Air Traffic Control </li></ul></ul></ul></ul>
    12. 12. Functional vs. Subject Oriented Taxonomies <ul><li>Data should be defined in a stand alone fashion, independent of function, in order for it to be useful to multiple functions or purposes </li></ul><ul><li>Having multiple taxonomies is acceptable but: </li></ul><ul><ul><li>Functional taxonomies should not lead to defining or structuring the data in a functional way </li></ul></ul><ul><ul><li>Each taxonomy that an organization creates must be maintained – including any necessary mapping - overhead </li></ul></ul><ul><li>Subject-oriented taxonomies offer: </li></ul><ul><ul><li>Potential for completeness </li></ul></ul><ul><ul><li>Stability </li></ul></ul>