Explainable AI Framework for Predicting Student Academic Success Through Personality Analysis

Abstract

The correlation between student personality traits and academic achievement has been a fundamental aspect of educational psychol ogy; yet, conventional analytical techniques frequently fall short in the prediction capability and interpretability required for practical appli cations. This research introduces an explainable artificial intelligence (XAI) framework that utilizes interpretable machine learning models to forecast student academic performance based on personality traits. We gathered data from 850 undergraduate students from various disciplines, including Big Five personality survey scores, demographic details, and cumulative academic performance markers. Various classification and regression models were developed and assessed, including Random For est, Gradient Boosting, and Neural Networks, utilizing SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-agnostic Explanations) methods to guarantee model transparency. Our research indicates that conscientiousness and openness to experience are the most significant predictors of academic achievement, while the explainability layer offers detailed insights into individual prediction trajectories. The suggested framework attained 87.3% accuracy in performance classifi cation while ensuring complete interpretability, allowing educators and administrators to identify at-risk students and formulate individualized intervention programs. This study illustrates how XAI can reconcile prediction accuracy with human comprehension in educational analyt ics, facilitating data-informed decision-making that upholds student pri vacy and advances equitable learning results.