A Method for Predicting Human Performance of Certain Symbol Recognition Tasks.
AIR FORCE INST OF TECH WRIGHT-PATTERSON AFB OHIO SCHOOL OF ENGINEERING
Pagination or Media Count:
The purpose of the study was to model human performance in a symbol recognition task. An algorithm is developed enabling one to closely model the results obtained from human subjects in the test. This model is then used to predict the human performance of another recognition test conducted by a different group of psychologists using three completely different symbol sets. The model has enabled the author to predict the results obtained from human testing with very high accuracy. It is concluded that the results may possibly reduce the necessity for further human testing of the symbol recognition task modeled by this algorithm. Modified author abstract