No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

The Refiner series of intelligent systems are used to create non-overlapping category descriptions from a set of cases which have been assigned to categories by an expert.  The systems generalise values for descriptors (fields) for each of the categories to create category descriptions, and suggest to the user various means by which any inconsistencies can be removed. In this thesis we define a new expert system, AKT-R4, which is the fourth iteration of the Refiner series.  Unlike the previous Refiner systems, the focus on AKT-R4 is classification rather than the creation of unambiguous category descriptions. The AKT-R4 system is based on a case-based algorithm focussed on performing a classification task, in particular medical (differential) diagnosis (a classic classification task); AKT-R4 is aimed primarily at medical students and junior doctors (i.e. novice diagnosticians) although it is not domain specific. The AKT-R4 algorithm incorporates concepts from case-based reasoning, hypothetico-deductive reasoning and illness scripts, and performs diagnosis by means of a new concept known as the diagnosis web, which is an expansion of the illness script concept.  This system supports the novice user as he/she develops and navigates through a medical knowledge web and requires no additional knowledge acquisition beyond the specification of a set of cases which the system uses to build a model of the domain. Medical diagnosis is difficult to perform and difficult to teach as it requires a large volume of information of various kinds, and an appreciation of the differential diagnosis process.  As such, an additional benefit of the AKT-R4 system is that it can be used to tutor diagnosis. A preliminary evaluation of the system has been performed and some useful feedback has been received.

Objective assessment of the neonatal outcome of labour is important, but it is a difficult and challenging problem. It is an invaluable source of information which can be used to provide feedback to clinicians, to audit a unit's overall performance, and can guide subsequent neonatal care. Current methods are inadequate as they fail to distinguish damage that occurred during labour from damage that occurred before or after labour. Analysis of the chemical acid-base status of blood taken from the umbilical cord of an infant immediately after delivery provides information on any damage suffered by the infant due to lack of oxygen during labour. However, this process is complex and error prone, and requires expertise which is not always available on labour wards. A model of clinical expertise required for the accurate interpretation of umbilical acid-base status was developed, and encapsulated in a rule-based expert system. This expert system checks results to ensure their consistency, identifies whether the results come from arterial or venous vessels, and then produces an interpretation of their meaning. This 'crisp' expert system was validated, verified and commercially released, and has since been installed at twenty two hospitals all around the United Kingdom. The assessment of umbilical acid-base status is characterised by uncertainty in both the basic data and the knowledge required for its interpretation. Fuzzy logic provides a technique for representing both these forms of uncertainty in a single framework. A 'preliminary' fuzzy-logic based expert system to interpret error-free results was developed, based on the knowledge embedded in the crisp expert system. Its performance was compared against clinicians in a validation test, but initially its performance was found to be poor in comparison with the clinicians and inferior to the crisp expert system. An automatic tuning algorithm was developed to modify the behaviour of the fuzzy model utilised in the expert system. Sub-normal membership functions were used to weight terms in the fuzzy expert system in a novel manner. This resulted in an improvement in the performance of the fuzzy expert system to a level comparable to the clinicians, and superior to the crisp expert system. Experimental work was carried out to evaluate the imprecision in umbilical cord acid-base parameters. This information, in conjunction with fresh knowledge elicitation sessions, allowed the creation of a more comprehensive fuzzy expert system, to validate and interpret all acid-base data. This 'integrated' fuzzy expert system was tuned using the comparison data obtained previously, and incorporated vessel identification rules and interpretation rules, with numeric and linguistic outputs for each. The performance of each of the outputs was evaluated in a rigorous validation study. This demonstrated excellent agreement with the experts for the numeric outputs, and agreement on a par with the experts for the linguistic outputs. The numeric interpretation produced by the fuzzy expert system is a novel single dimensional measure that accurately represents the severity of acid-base results. The development of the crisp and fuzzy expert systems represents a major achievement and constitutes a significant contribution to the assessment of neonatal outcome.

No description available.

No description available.