Concept Trees: Building Dynamic Concepts from Semi-structured Data Using Nature-Inspired Methods - Complex System Modelling and Control Through Intelligent Soft Computations - page 223

information that they typically use. This would naturally happen as part of the
Semantic Web. The construction of these ontologies will enable computers to
autonomously search the Internet and interact with the services that are provided
and is also part of knowledge management on the Internet. The topic

' Towards the

Semantic Web: Ontology-Driven Knowledge Management ( 2003 ) discusses the

ontology construction problem in relation to p2p networks and the Semantic Web.
They note that for reasons of scalability, ontology construction must be automated,
based on information extraction and natural language processing technologies.
However, for reasons of quality, the process still requires a human in the loop, to
build and manipulate ontologies. With a slightly reduced knowledge-level, it is
intended that the concept base can construct itself almost completely autonomously,
giving it a major advantage in this respect.
For dynamic or autonomic systems, the context in which the knowledge is used
can become a critical factor. Context is an information space that can be modelled
as a directed graph, rather like an ontology. Context allows both recognition and
mapping of knowledge, by providing a structured and uni

'

ed view of the world in
which it operates (Coutaz et al. 2005 ). It is about evolving, structured and shared
spaces that can change from one process to the next, or even through the duration of
a single process. As such, the meaning of the knowledge will evolve over time. The
key lies in providing an ontological foundation, an architectural foundation, and an
approach to adaptation that all scale alongside the richness of the environment.
Contexts are de

c set of situations, roles, relations and entities.
A shift in context corresponds to a change in the set of entities, a change in the set
of possible relations between entities, or a change in the set of roles that entities
may play. Unfortunately, the context adaptation cannot currently be carried out in a
totally automatic way and a concept base would not really consider context in the

ned by a speci

first instance. It is constructed primarily through statistical counts, but groups of
terms presented at the same time can provide some level of context.
By describing the domain in a standardised way, the programs that use the
domain will be able to understand what the domain represents. Through this pro-
cess, different programs on the Internet will be able to learn about each other and
form useful associations with other programs that provide the information that they
require. This will enrich the knowledge that they can provide, thus turning the
Internet into a knowledge-based system, rather than primarily as a source for direct
information retrieval. This is of course, a utopian idea that has many possibilities
and may never be fully realised.

3.2 Dynamic Databases

As a concept base is a type of database, this is probably the first technology to look

at, where the following text is also taken from the topic Greer ( 2008 , Chap. 3).
Databases are the

first models
were developed in the 1960s. The relational model proposed by Codd ( 1970 ) has

first kind of organised information system, where the