Semantic Modeling for Data: Avoiding Pitfalls and Breaking Dilemmas

Cover
Copyright
Table of Contents
Preface
	Who Should Read This Book
	What to Expect in This Book
	Book Outline
	Conventions Used in This Book
	O’Reilly Online Learning
	How to Contact Us
	Acknowledgments
Part I. The Basics
	Chapter 1. Mind the Semantic Gap
		What Is Semantic Data Modeling?
		Why Develop and Use a Semantic Data Model?
		Bad Semantic Modeling
		Avoiding Pitfalls
		Breaking Dilemmas
	Chapter 2. Semantic Modeling Elements
		General Elements
			Entities
			Relations
			Classes and Individuals
			Attributes
			Complex Axioms, Constraints, and Rules
			Terms
		Common and Standardized Elements
			Lexicalization and Synonymy
			Instantiation
			Meaning Inclusion and Class/Relation Subsumption
			Part-Whole Relation
			Semantic Relatedness
			Mapping and Interlinking Relations
			Documentation Elements
		Summary
	Chapter 3. Semantic and Linguistic Phenomena
		Ambiguity
		Uncertainty
		Vagueness
		Rigidity, Identity, Unity, and Dependence
		Symmetry, Inversion, and Transitivity
		Closed- and Open-World Assumptions
		Semantic Change
		Summary
	Chapter 4. Semantic Model Quality
		Semantic Accuracy
		Completeness
		Consistency
		Conciseness
		Timeliness
		Relevancy
		Understandability
		Trustworthiness
		Availability, Versatility, and Performance
		Summary
	Chapter 5. Semantic Model Development
		Development Activities
			Setting the Stage
			Deciding What to Build
			Building It
			Ensuring It’s Good
			Making It Useful
			Making It Last
		Vocabularies, Patterns, and Exemplary Models
			Upper Ontologies
			Design Patterns
			Standard and Reference Models
			Public Models and Datasets
		Semantic Model Mining
			Mining Tasks
			Mining Methods and Techniques
		Summary
Part II. The Pitfalls
	Chapter 6. Bad Descriptions
		Giving Bad Names
			Setting a Bad Example
			Why We Give Bad Names
			Pushing for Clarity
		Omitting Definitions or Giving Bad Ones
			When You Need Definitions
			Why We Omit Definitions
			Good and Bad Definitions
			How to Get Definitions
		Ignoring Vagueness
			Vagueness Is a Feature, Not a Bug
			Detecting and Describing Vagueness
		Not Documenting Biases and Assumptions
			Keeping Your Enemies Close
		Summary
	Chapter 7. Bad Semantics
		Bad Identity
			Bad Synonymy
			Bad Mapping and Interlinking
		Bad Subclasses
			Instantiation as Subclassing
			Parts as Subclasses
			Rigid Classes as Subclasses of Nonrigid Classes
			Common Superclasses with Incompatible Identity Criteria
		Bad Axioms and Rules
			Defining Hierarchical Relations as Transitive
			Defining Vague Relations as Transitive
			Complementary Vague Classes
			Mistaking Inference Rules for Constraints
		Summary
	Chapter 8. Bad Model Specification and Knowledge Acquisition
		Building the Wrong Thing
			Why We Get Bad Specifications
			How to Get the Right Specifications
		Bad Knowledge Acquisition
			Wrong Knowledge Sources
			Wrong Acquisition Methods and Tools
		A Specification and Knowledge Acquisition Story
			Model Specification and Design
			Model Population
		Summary
	Chapter 9. Bad Quality Management
		Not Treating Quality as a Set of Trade-Offs
			Semantic Accuracy Versus Completeness
			Conciseness Versus Completeness
			Conciseness Versus Understandability
			Relevancy to Context A Versus Relevancy to Context B
		Not Linking Quality to Risks and Benefits
		Not Using the Right Metrics
			Using Metrics with Misleading Interpretations
			Using Metrics with Little Comparative Value
			Using Metrics with Arbitrary Value Thresholds
			Using Metrics That Are Actually Quality Signals
			Measuring Accuracy of Vague Assertions in a Crisp Way
			Equating Model Quality with Information Extraction Quality
		Summary
	Chapter 10. Bad Application
		Bad Entity Resolution
			How Entity Resolution Systems Use Semantic Models
			When Knowledge Can Hurt You
			How to Select Disambiguation-Useful Knowledge
			Two Entity Resolution Stories
		Bad Semantic Relatedness
			Why Semantic Relatedness Is Tricky
			How to Get the Semantic Relatedness You Really Need
			A Semantic Relatedness Story
		Summary
	Chapter 11. Bad Strategy and Organization
		Bad Strategy
			What Is a Semantic Model Strategy About?
			Buying into Myths and Half-Truths
			Underestimating Complexity and Cost
			Not Knowing or Applying Your Context
		Bad Organization
			Not Building the Right Team
			Underestimating the Need for Governance
		Summary
Part III. The Dilemmas
	Chapter 12. Representation Dilemmas
		Class or Individual?
		To Subclass or Not to Subclass?
		Attribute or Relation?
		To Fuzzify or Not to Fuzzify?
			What Fuzzification Involves
			When to Fuzzify
			Two Fuzzification Stories
		Summary
	Chapter 13. Expressiveness and Content Dilemmas
		What Lexicalizations to Have?
		How Granular to Be?
		How General to Be?
		How Negative to Be?
		How Many Truths to Handle?
		How Interlinked to Be?
		Summary
	Chapter 14. Evolution and Governance Dilemmas
		Model Evolution
			Remember or Forget?
			Run or Pace?
			React or Prevent?
			Knowing and Acting on Your Semantic Drift
		Model Governance
			Democracy, Oligarchy, or Dictatorship?
			A Centralization Story
		Summary
	Chapter 15. Looking Ahead
		The Map Is Not the Territory
		Being an Optimist, but Not Naïve
		Avoiding Tunnel Vision
		Avoiding Distracting Debates
			Semantic Versus Nonsemantic Frameworks
			Symbolic Knowledge Representation Versus Machine Learning
		Doing No Harm
		Bridging the Semantic Gap
Bibliography
Glossary
Index
About the Author
Colophon