Page 145 -

128 Chapter 4

Table 4.2
Major taxonomic approaches to knowledge codifi cation

Taxonomic approach Key features
Cognitive or concept map • Each key content item is represented as a node in a
graph and the relationships between these key concepts
are explicitly defi ned.
• Can show multiple perspectives or views on the same
content.
• Fairly easy to produce and intuitively simple to
understand but diffi cult to use for knowledge related
procedures.
Decision tree • Hierarchical or fl owchart type of representation of a
decision process.
• Very well suited to procedural knowledge — less able to
capture conceptual interrelationships.
• Easy to produce and easy to understand.
Manual knowledge • Object-oriented approach that allows lower or more
taxonomy specifi c knowledge to automatically incorporate all
attributes of higher-level or parent content they are
related to.
• Very fl exible — can be viewed as a concept map or as a
hierarchy.
• More complex, therefore will require more time to
develop, as they must refl ect user consensus.
Automated knowledge • A number of tools are now commercially available for
taxonomy taxonomy construction.
• Most are based on statistical techniques such as cluster
analysis to determine which types of content are more
similar to each other and can constitute subgroups or
thematic sets.
• Good solution if there is a large amount of legacy
content to sort through.
• More expensive and still not completely accurate — will
need to be validated and refi ned for maximum usefulness.