Related Experiment Videos
Contextual pathognomony: a computationally useful extension of pathognomony
1Laboratory for Knowledge Based Medical Systems, Ohio State University, Columbus.
Computer Methods and Programs in Biomedicine
|August 1, 1991
Summary
Pathognomonic findings offer focus in expert systems but are rare. This study introduces contextual pathognomony, a more computationally useful concept, demonstrated in alloantibody identification.
Area of Science:
- Artificial Intelligence
- Medical Informatics
Background:
- Abductive expert systems often struggle with problem-solving focus.
- Pathognomonic findings, while ideal for focus, are too rare for practical use.
- Pople's 'constrictor' concept offers a more computationally useful alternative by broadening focus to disease classes.
Purpose of the Study:
- To introduce and define the concept of contextual pathognomony.
- To demonstrate the computational utility of contextual pathognomony in abductive reasoning.
- To validate the concept in the domain of alloantibody identification.
Main Methods:
- Conceptual extension of pathognomony to include diagnostic context.
- Experimental analysis of contextual pathognomony's effectiveness.
- Application within the specialized domain of alloantibody identification.
Main Results:
- Contextual pathognomony is more ubiquitous and computationally useful than traditional pathognomony.
- The concept effectively provides focus for abductive reasoning.
- Experimental evidence supports the utility of contextual pathognomony in alloantibody identification.
Conclusions:
- Contextual pathognomony offers a practical solution to the focus problem in abductive systems.
- This concept enhances computational efficiency in diagnostic reasoning.
- The findings have implications for developing more effective expert systems in medicine.