Exploring The Use of Explainable Ai (Xai) For Personalized Educational Feedback Systems

Abstract

Explainable AI (XAI) plays a crucial role in meeting the demand for greater transparency and interpretability in AI-driven decision-making within the context of personalized educational feedback systems. The research introduces a novel model called XAI-LSTM-PEF (Explainable AI-based Long Short-Term Memory Personalized Educational Feedback) that aims to produce personalized, transparent, and actionable feedback for learners. The novelty of the proposed model is that it cannot only predict students' learning outcomes, but also give students explanatory information about what factors affect the prediction, which enhances the trust and participation of students. The use of LSTM networks for sequential data analysis, along with XAI methods such as LIME and SHAP, allows for the generation of interpretable feedback, thus meeting the demand for personalization and transparency. Experimental results show that the XAI-LSTM-PEF model outperforms the traditional models, including Neural Network (NN) and Logistic Regression (LR) with an accuracy of 92%, a precision of 94% and an F1 score of 97%. Moreover, the model exhibits a high ROC AUC of 95% for the evaluation, which depicts its good performance on various evaluation measures. These results demonstrate the potential of XAI in fostering a more personalized and transparent learning environment, ultimately improving educational outcomes and fostering student trust in AI systems.