A Model to Represent Knowledge about Assistive Products

Abstract

This paper describes a model which has been developed to represent knowledge about assistive products, to feed an artificial-intelligence-based online system offering guidance to identify and select the assistive products that best suit individual needs. In this model, each assistive product is described by a set of “knowledge rules” clustered round 15 chapters: 1) product identification data and overall description; 2) possible configuration variants; 3) optional components; 4) product goals; 5) indicated impairments and 6) contraindicated impairments; 7) indicated and 8) contraindicated environments; 9) other indicated and 10) contraindicated factors; points to consider in 11) selection, 12) fitting, 13) use and 14) maintenance/follow-up; and 15) sources/references. Each “knowledge rule” consists of a sentence – written in English language according to given guidelines – each containing a token of knowledge provided by an expert, based on scientific evidence or field experience; in this way, the knowledge base grows token by token thanks to the collective effort of a worldwide community of experts, each entering their own tokens on a voluntary basis. Today, the knowledge base includes about 2400 knowledge rules, mainly related to products belonging to the WHO APL (Assistive Product Priority List). It feeds an online guidance system called “Assistive Product Explorer” (ASPREX) which is currently under development by the World Health Organization within the GATE initiative (Global Collaboration on Assistive Technology). The model has shown able to represent knowledge about any categories of assistive products, and suitable for being fed by an open community of experts worldwide through the ASPREX system.