Semantic modelling of product support Knowledge

R.M. Setchi, N. Lagos, S.S. Dimov

Cardiff University, Cardiff , UK

Introduction

• Companies offer various forms of assistance
• Supplies, tools, equipment, information
• Knowledge is a critical part of any pro-active and creative business strategy but product support has not been viewed as a knowledge intensive process

• Aim: Model product support based on knowledge engineering methodology

Background and related work

• Product support is "everything necessary to allow the continued use of a product" (Pham et al. 1999)

• Interactive Electronic Technical Manuals (IETMs)
• Intelligent Product Manuals (IPMs)
• Electronic Performance Support Systems (EPSSs)

• Integration of knowledge engineering practices in product support systems
• Reasoning mechanisms
• Knowledge base
• Semantic data modelling

• Previous work does not follow a uniform approach towards the development of knowledge bases →Ontological principles are utilised in product support knowledge modelling

Knowledge modelling approach

• Product support knowledge sets
• Product knowledge
• Task knowledge
• User knowledge
• Data to Knowledge
• Raw data is gathered
• Common relations among data items are identified →Organised information
• Developed structures are further analysed to make them available for productive use

Knowledge modelling approach-continued

• Conceptual modelling → Formalise domain knowledge in a high-level, yet authoritative way

• First phase: Objective is to organise data in a way that ensures homogeneity and validity of the resulting information
• Use of ontological principles and KE notions: Concept, Instance, Relation
• Creation of new structures: Domain, connector
• The architectural model is developed at the end of the first phase

• Second phase: Objective is to put the structured information into productive use, within the context of product support
• Architectural model is enriched with two notions: Knowledge-specifier and arc
• Connectors are characterised as active or inactive: Connector-dependence measure
• The functional model is developed at the end of the second phase

Implementation

• The knowledge base is developed using the knowledge model described above, populated with instances and their associations
• The Protege ontology editor has been used
• Currently, the knowledge base contains 278 frames, including 75 concepts, 80 slots, 22 facets, and 101 instances and part of it is depicted in the next slide

Implementation-continued