Explainable AI and machine learning: performance evaluation and explainability of classifiers on educational data...

Area of Science:

• Artificial Intelligence
• Educational Technology
• Data Science

Background:

• Machine learning (ML) and Explainable AI (XAI) offer advanced capabilities for analyzing complex data.
• Educational frameworks can leverage ML/XAI to provide intelligent support for students' career development.
• Current career counseling methods can be enhanced by data-driven insights into academic and employability factors.

Purpose of the Study:

• To propose an ML and AI-powered framework for student career counseling.
• To analyze educational datasets for factors influencing career placement and growth.
• To integrate expert system functionalities for decision support in career choices.

Main Methods:

• Development of a framework integrating ML and XAI for educational analysis.
• Utilization of ML-based White Box and Black Box models.
• Training models on an educational dataset encompassing academic and employability attributes.

Main Results:

• Naive Bayes models demonstrated superior performance in prediction tasks.
• Achieved a Recall score of 91.2% and an F-Measure score of 90.7% using Naive Bayes.
• Outperformed Logistic Regression, Decision Tree, SVM, KNN, and Ensemble models in the study.

Conclusions:

• The proposed framework effectively utilizes ML and AI for career counseling.
• Naive Bayes proved to be the most effective model for predicting student career outcomes within the study's dataset.
• The framework provides valuable decision support for students' career planning and placement.