Page 170 -

8 The Importance of Ontological Structure: Why Validation by ‘Fit-to-Data’... 167

Table 8.4 Comparison of expressivity of ontologies of various modelling approaches
Modelling
approach Expressions needed Description logic Comments
Neural The concepts of inputs AL .D/ The only ontologically
networks and outputs and data significant terms are the input
property labels for each and output variables.
node Rudimentary classes are
needed for input and output
specifically
ODEs Data properties for each ALCOIN .D/ Concepts would be needed for
variable, distinction each variable so that causal
between exogenous and influences can be represented
endogenous variables, with relationships
causal influence
System As ODEs, but stocks ALCOIN .D/ Stocks and flows as concepts
dynamics and flows are also do not add any extra
relevant concepts requirements for expressivity
Social Individuals and ALI .D/ Concepts not really needed
network relationships, possibly (other than Top), so ALI is
analysis data properties where more expressive than SNA
attributes of individuals really requires. Data properties
relevant optional
Agent-based Classes, inheritance, ALCROINF .D/ Not all agent-based models will
modelling individuals, data need all the expressivity
properties, object options. If you have a NetLogo
properties, lists, arrays, model and want to find out the
domain and ranges expressivity of its ontology,
needed you can use Polhill’s (2015)
automated ontology extraction
tool and load the result into
Protégé, and the ontology
summary tab tells you the
description logic needed. For
example, Ge and Polhill’s
(2016) model of commuting
has description logic
.D/
ALRIF
Letters are used to represent terms or groups of terms needed to capture any syntax for the
modelling approach’s formalism with respect to the real world. See text for an explanation

in Java); I inverse relationships; N numerical restrictions on properties; and F
functional properties. Table 8.4 provides an initial indication of the description logic
expressivity needed to capture the syntax used to specify the ontologies of various
modelling approaches. However, the labels applied in the ‘description logic’ column
do not necessarily mean that the full capabilities of the language are necessarily
used.