A Development of Fuzzy Logic Expert-Based Recommender System for Improving Students' Employability

Most of the higher education institutions in the world have already been evaluating their strategies of enhancing the employability of their students and introducing different ways to improve and strengthen it. Identifying student's employability and recommend the areas for improvement before graduation will increase the chance to be employed if the students developed their employability skills. That is why in this paper the rule-based algorithm that is commonly used in developing a recommender system were replaced by Fuzzy logic based because rule-based cannot eliminate the ambiguity issues in decision-making, whether they are hirable or not before graduation. Various factors may affect the employability of undergraduate students. In this paper, the employability prediction and recommender system for students were built using fuzzy logic to resolve the issue. The most significant attributes that affect the undergraduate students' employability were determined using feature selection filtering techniques and used as inputs. The result shows that the developed fuzzy model performs a high predictive accuracy based on the computed mean absolute error (MAE) and root-mean-square error (RMSE) scores which decrease from the training to the validation and test sets.