ONTOCOPI uncovers Communities Of Practise by analysing the connectivity of instances in the 3store knowledge base, shown here running as a plugin to Protégé 2000

ONTOCOPI fact-file

What's the Problem?

Communities of practice (COPs) are informal self-organising groups of individuals interested in a particular practice (special interest groups, expertise networks, etc). Identifying COPs is often regarded as an essential first step to understanding the knowledge resources of an organisation. COP identification is currently a resource-heavy process largely based on interviews that can be very expensive and time consuming, especially if the organisation is large or distributed.

Towards a Solution

Ontocopi (Ontology-based Community of Practice Identifier) attempts to uncover COPs by applying a set of ontology-based network analysis techniques that examine the connectivity of instances in the knowledge base with respect to the type, density, and weight of these connections.

The advantage of using an ontology to analyse such networks is that relations have semantics or types. Hence certain relations - the ones relevant to the COP - can be favoured in the process of analysis. During the analysis the weight of the contribution made by the important relations is high, while that of the less important ones can be made relatively low, or zero.

Ontocopi applies an expansion algorithm that generates the COP of a selected instance (could be any type of object) by identifying the set of close instances and ranking them according to the weights of their relations. It applies a breadth-first, spreading activation search, traversing the semantic relations between instances (ignoring directionality) until a defined threshold is reached.

Consider the example network on the right. Assume we need to identify the COP of the query instance A (which could be a person, project, document, etc), using the relationships hasAuthor, memberOf and attended, with the weights 1.0, 0.6, and 0.3 respectively. All instances will have an initial weight of 1. Activation will spread from the query instance to neighbouring instances in the network, up to a given number of links. In the first expansion, the query instance A will pass on weight to all the instances it is connected to. The weight passed equals the weight of the instance multiplied by the weight of the traversed relationship. In this case, A passes 1*0.6 to D, and 1*1 to H. These will be added to their initial weights of 1. In return, these instances will pass their total weights to all their neighbours, so D for example will pass (1+1*0.6)*0.6 to B and A. Expansion will stop when the link paths are exhausted or the link threshold is reached. Instances therefore accumulate weight based on the number of relevant relations they have with the initial instance.

Ontocopi currently exists in three different implementations;

1. As a plugin to Protégé 2000 (pictured above), with an exemplar knowledge base of around 13000 instances taken from the ECS department at the University of Southampton. The plugin offers many features that can user can modify to control the type and scope of the required COP.

2. As an applet working with the 3store knowledge base, which is a large repository of RDF data about UK computer scientists.

3. As a standalone application than can be queried online and returns COPs in RDF
   This query will return the COP of Harith Alani

Take a Guided Tour

Here is a brief demo of Ontocopi Protégé's plugin (Quick Movie, Avi).

Try a Demonstration

ontocopi applet (beta version)

Technical requirements: java enabled browser.

Semantic representation

View in the AKT Triplestore Browser or as RDF.

Also available in DOAP RDF (Description Of A Project)