Knowledge Acquisition/Engineering related concepts
As understood by
Philippe A. MARTIN
- Knowledge acquisition/engineering (KA/KE)
- Goal: Building a knowledge-based system (KBS)
- Method:
- Prototype-based knowledge acquisition/engineering
- Goal: Building a knowledge-based system (KBS)
- Subtask: (Adapted from: Heijst & al., 1996)
- Data collection
- Input: Expertise sources
- Output: Expertise data (e.g.: organisational model,
interview retranscription, technical documents)
- Prototype design (Input: Expertise data; Output: KBS)
- Method:
- Using ripple-down procedures (Paul Compton)
-
- Model-based knowledge acquisition/engineering
- Goal: Building a knowledge-based system (KBS)
- Subtask: (Adapted from: Heijst & al., 1996)
- Data collection
- Knowledge modelling
- Update/complement the library/ontology of generic models
- Knowledge implementation
- Subtask:
- Knowledge modelling (Adapted from: Heijst & al., 1996)
- Subtask:
- Output:
- Conceptual model
- Part:
- Task model
(Part: Application task ontology,
Application task knowledge)
- Domain model
(Part: Application domain ontology,
Application domain knowledge)
- Ressource:
- Generic models/ontology
- Part: Generic task models/ontology,
Generic domain models/ontology
- Building the application domain knowledge
- Subtask:
- Knowledge base validation
- Subtask:
- Knowledge implementation validation
- Subtask: (Adapted from: Heijst & al., 1996)
- Showing the reasoning steps at different levels of abstraction (as in: Krest, Dids)
- Exploiting the reasoning trace to answer Why and How questions (as in: Kew)
- Simulation for answering What_if and How_to questions (as in: Salt)
- Finding the source of a reasoming mistake (as in: Mole)