Fuzzy Lattice Reasoning for Intelligent Student Performance Evaluation in Educational Settings

In this paper, a novel approach to the allocation of students to performance levels – Good, Average,and Weak, using fuzzy lattice logic, is proposed. The traditional grading system tends to establishhard boundaries and cutoffs for allocation, whereas student performance is fuzzy. In view of this,a fuzzy lattice classifier able to effectively manage fuzziness with a logical structure for allocationof students has been proposed.We utilize an extensive palette of academic metrics and develop a partially ordered structure onwhich fuzzy inclusion defines the formation of student groups. Applying the method developed toa set of 450 university students, the accuracy of student classification was found to be 91.3%,exceeding the accuracy of traditional classifiers such as decision trees of 84.7% and neuralnetworks of 86.2%. The method has equally improved on interpretability.