Flexible, adaptive, automatic fuzzy-based grade assigning system

摘要

In assigning course grades at the end of a term, most instructors are faced with the task of converting a collection of numerical scores into a single letter grade. This conversion process is generally accomplished in two steps: (1) the collection of numerical scores obtained over the course of a grading period is used to compute a weighted average; and (2) the final term average is used to assign a letter grade. The paper addresses the second step in the grade assignment process. The development and testing of a flexible, adaptive, automatic grading system is discussed. The grading system is based on fuzzy mathematics which allows the spreadsheet based grader to model the subjective nature of human grade assigners. A genetic algorithm is then employed to provide the grading system with the adaptive capabilities necessary to alter grade assigning strategies in concert with the attitudes of a particular instructor.