Upload
ajrhem
View
722
Download
1
Tags:
Embed Size (px)
DESCRIPTION
Presentation on Taxonomies and Ontology\'s for Enterprise Search (Includes 3 case study examples)
Citation preview
Company
LOGO
Designing and Implementing Taxonomies and Ontology's in
Enterprise Search
KM World Taxonomy Boot Camp - 2011
Overview
Overview: looking at how several organizations use taxonomies and ontology's to improve unstructured content search and retrieval and meet the business expectations of the KM solution
3
Objectives
• This presentation will focus on the design and implementation of taxonomies and ontology's to improve unstructured content search and retrieval. Specifically this presentation will take a look at several organizations and how the approach to enterprise search enables a successful search result. Along with presenting examples of taxonomy adoption an underlying view of the content types and metadata will be presented that met the business expectations of the KM solution.
4
Agenda
• Taxonomy and Ontology• Information Model• Unstructured Data• Content Types and Metadata• Search Engines
• Microsoft Fast• Google Search Appliance
• Case Studies• Military Organization• Retail Organization• Financial Organization
Taxonomies and Ontology's
Taxonomy: the science or technique of classification; a classification into ordered categories; example a taxonomy of animals.
Taxonomies and Ontology's are a way of classifying something
5
Ontology: ontology deals with questions concerning what entities exist or can be said to exist, and how such entities can be grouped, related within a hierarchy, and subdivided according to similarities and differences; example a ontology of a car.
6
Information Model (Facts)
• Information Model is typically developed from the Ontology• Business Rules around the information relationships are
established• The Business Rules contributed to the construction of the
information model• The information represents a sharable, stable, and
organized structure of information requirements for your Knowledge Management System (KMS)
• Information Model supports the search process through establishing relationships between the content and describing how this information behaves
Information Model Example
Source: CMBL Information Model - http://www.mod.uk/NR/rdonlyres/C176E21A-776C-46FA-AE2B-3CAD597CDD6A/0/CBML_information_model.pdf
8
• Unstructured Data – In contrast to structured data, unstructured data has no identifiable structure associated with it
• Unstructured data comes in the form of:• Images/Objects• Email• Documents (word, PDF, etc.)• Spreadsheets (i.e., excel)
• Most data in the enterprise today is in the form of unstructured data• Unstructured data contains the explicit knowledge of the enterprise and
has to be made available to the knowledge management system• In order to catalog, search and retrieve unstructured data we must
make it identifiable by building structure around it.• This structure comes in the form of Content Types and Metadata.
Unstructured Data
9
• Content Types – a reusable collection of metadata, and other settings for a category of artifacts, items, or documents
• Content types enable you to manage the settings for a category of information in a centralized and reusable way
• Content Types encapsulate data requirements
• Content Types enable Content Standardization and are File Format Independent
• Metadata – The metadata represents the properties of a Content Type
Content Types and Metadata
Content Type: ApplicationColumn Name TypeApplication Name ChoiceDescription TextOwner LookupVendor ChoiceType Choice
Search Engines
Microsoft FAST for SharePoint Provided: Directly index against the content Advance Filtering Navigation breadcrumbs Unsupervised clustering Concept Extraction
10
Google Search Appliance (GSA) Provided: Dynamic Scalability - Scale to millions of documents/artifacts Fine tune relevancy - Ranking Framework, Node Biasing, and Collection Biasing Customizable security, enabling early binding and late binding Social search features, including 'User Added Results' User-centric innovations such as Query Suggestions Enhanced search quality with improved precision
The Following Case Studies utilized either Microsoft SharePoint Search, Microsoft Fast for SharePoint or Google Search Appliance (GSA):
11
Case Study – Military Organization
• Opportunity: Capture of Tacit and Explicit Knowledge (retiring and rotational workforce) and rules in response to Defense Base Closure and Realignment (BRAC) Commission movements
• Activities:- Knowledge Capture- Create Knowledge Repository- Implement Enterprise Search
• Results:- Knowledge Identified/Cataloged (Key Knowledge Loss Avoided)- Utilized Taxonomy to Structure the Site and Categorize the Content- Utilized Ontology/Information Model to establish information
relationships and contribute to search engine optimization
12
Case Study – Military Organization – Taxonomy/Content Structure
Taxonomy: Provided infrastructure to deliver Site and Content structure
Taxonomy StructureDirectorate-- Division---- Groups------ Battalions
Content Structure within the Taxonomy/SOP/Training/Projects/Plans/General Admin/Policy and Procedures
13
Case Study – Military Organization – Site Structure (1)
14
Case Study – Military Organization – Site Structure (2)
15
Case Study – Military Organization – Ontology/Info Model
Ontology/Information Model: Capture the information relationships and contribute to Search Engine Optimization
Personnel
Directorate
Division
System
Battalions
Groups
Readiness & Mobilization
Commands 1 or More
Support Operations
Support OperationsSupport Operations
Support OperationsCommands 1 or More
Commands 1 or More
Is a Kind ofIs a Kind of
Role
Performs duties within
Performs duties withinPerforms duties within
Performs duties within
Chief of Staff
Commanding General
BRAC POC Director Sealift Operations
Is a Kind ofIs a Kind of Is a Kind of
Is a Kind of
16
Case Study – Military Organization - Search
• Search decision - Recommended the Use of Google Search Appliance (GSA) to Provide:
• Dynamic Scalability• Fine Tune Relevancy• Customizable Security• Social Search Features• User-centric functionality • Enhanced Search Quality
• However; initial implementation utilized SharePoint out-of-the-box search capabilities with future enhancements to consider GSA or Microsoft Fast.
17
Case Study – Retail Organization
• Opportunity: Capture of Tacit and Explicit Knowledge of Vendors and make this knowledge available to associates. Lessen the need for company SME’s and enable vendor knowledge transfer.
• Activities:- Development of Taxonomy; Information Model; and Content
Types/Metadata- Performed Vendor Knowledge Capture- Create Knowledge Repository
• Results:- Knowledge Identified/Cataloged (Key Vendor assets Captured)- Established a standardized processes for capturing, storing, and
searching intellectual assets- Software Project Ramp up time decreased- Improved utilization of SME’s
18
Case Study – Retail Organization – Taxonomy
19
Case Study – Retail Organization – Site Structure
Organizational Taxonomy
Organizational (level 2) Taxonomy
20
Case Study – Retail Organization – Ontology/Info Model
Results:- Knowledge Identified/Cataloged (Vendor Knowledge Cataloged)- Architecture will aid in fulfilling search requirements - Established Rules and Policies concerning information
Walmart Artifact
ISDLC Artifact
Division Business Unit
Product Family
Product Application
Country
KnOD
Consists of
Consists ofConsists of
Project
Contains
Has a Collection of
Has a Collection of
Belongs To
Is a Part of Can Be Associated to
Can Be Associated to
Can Be Associated
toCan Be
Associated to
Can Be Associated
to
Is Associated to
Categorizes items by
21
Case Study – Retail Organization – Content Types/Metadata
Content Types/Metadata: Will aid in the storing, and searching of Intellectual assets
Content Type: Company ArtifactMetadata Fields: Artifact Category
Artifact ContactConfidentiality Level (Shared, Controlled, or Restricted)SummaryLanguageSearch KeywordsCountryDivision
Content Type: ISDLC ArtifactMetadata Fields: Artifact Type
Project Id
22
Case Study – Retail Organization - Search
• Search decision - utilized SharePoint out-of-the-box search capabilities
• Although the initial implementation utilized SharePoint out-of-the-box search capabilities; future enhancements will implement Microsoft Fast for search. To provide the following search functionality:
Directly index against the contentAdvance FilteringNavigation breadcrumbs
23
Case Study – Financial Organization
• Opportunity: Transition, Capture and Catalog Tacit and Explicit Knowledge from across several business units and produce content that is solution base, fast and easily searchable and retrievable.
• Activities:- Provide Content Management- Provide Business Process Integration with Workflows- Establish Enterprise Search- Provide Admin and Business Intelligence Capabilities
• Results:- Knowledge Identified/Cataloged (Content Structured and Migrated)- Enterprise Search Enabled (Producing Solution Based Results)- Knowledge Portal Completed with BI, and Workflows Implemented
Content Type &Metadata Structure
MS Share Point 2010 Platform
KM Search Flow & Display
Work Flow (Operational
& Governance)
Present Content
Business Taxonomy
Migration Ready Content
Reporting
System Add – On (As Needed)
KM Enterprise Solution
Case Study – Financial Organization – KM Framework
25
Case Study – Financial Organization - Taxonomy
Taxonomy: Provided logical Site Structure and Content Structure for capturing and cataloging content for search.
Case Study – Financial Organization – Site Structure (1)
Case Study – Financial Organization – Site Structure (2)
28
Case Study – Financial Organization – Ontology/Info Model
Ontology/Information Model: Capture the information relationships and contribute to Search Engine Optimization
Account
Product
Procedure
Policy
Form
Business Unit
Division
Department
System
Client
Opens
Trade
Executes
Contains
Service
Supports the execution
Initiates
Futures Options Stock
Is a Kind of Is a Kind of Is a Kind of
Is a Kind of
1…*
0...1
Associated Policies
Associated Procedures
Contains
Contains Describes the use of
Describes the use of
Supportsthe
execution
Governs
1
1..,*
Governed byEstablishes
Initiates
FAQ
AnswersQuestions
about
Initiates
Initiates
Establishes
Establishes
Annuities
Mutual Funds
Exchange Traded Funds
Is a Kind of
Is a Kind of
Answers Questions About
RetailBrokerage Operations
Institutional
Services
Is a Kind of
Is a Kind of
Is a Kind of
Manages
Answers Questions About
Answers Questions About Answers
QuestionsAbout
29
Case Study – Financial Organization – Content Types/Metadata
Content Type Structure for Page Layout to capture web based Content
30
Case Study – Financial Organization – Content Types/Metadata
Content Type Structure for Documents to capture document (PDF, Excel, Word, etc.) based Content
Document
NameTitle
Business AreaRetail
Business AreaCorporate
Business AreaBrokerage Ops
Business AreaInstitutional
Is an occurrence of
Is an occurrence of
Is an occurrence of
Is an occurrence of
TDA-Artifact
Business AreaDivisionDepartmentArtifact TypeArtifact AliasClassOrderFamilyDescriptionOwnerPublish DateExpiration DateReview PeriodSecurity LevelKeywords
Forms
AliasDescriptionForm NumberRevision DateAffected SegmentsFaxableClient FacingFunctional CategorizationKeywords
31
Case Study – Financial Organization - Search
• Search decision - Utilized Microsoft Fast for SharePoint
• Microsoft Fast for SharePoint provided the following search functionality:
Directly index against the contentAdvance FilteringNavigation breadcrumbsUnsupervised clusteringConcept Extraction
32
Designing Taxonomies and Ontology's for Enterprise Search
33
Designing Taxonomies and Ontology's for Enterprise Search
A.J. Rhem & Associates, Inc.A.J. Rhem & Associates, Inc.500 North Michigan Ave., 500 North Michigan Ave., Suite 300Suite 300Chicago, Illinois 60611Chicago, Illinois 60611Phone: 312-396-4024Phone: 312-396-4024email: email: [email protected]@ajrhem.comWebsite: www.ajrhem.com