A pollen identification expert system ; an application of expert system techniques to biological identification : a thesis presented in partial fulfilment of the requirements for the degree of Master of Science in Computer Science Massey University
The application of expert systems techniques to biological identification has been investigated and a system developed which assists a user to identify and count air-borne pollen grains. The present system uses a modified taxonomic data matrix as the structure for the knowledge base. This allows domain experts to easily assess and modify the knowledge using a familiar data structure. The data structure can be easily converted to rules or a simple frame-based structure if required for other applications. A method of ranking the importance of characters for identifying each taxon has been developed which assists the system to quickly narrow an identification by rejecting or accepting candidate taxa. This method is very similar to that used by domain experts.