همایش بین المللی هوش مصنوعی

صفحه اصلی / دومین همایش بین المللی هوش مصنوعی

Development and Validation of an Explainable Machine Learning Framework for Decision Support in Student Admission Management

نویسندگان :

Hananeh Teshnehlab¹ Mohammad Reza Ayatollahzadeh Shirazi²

1- شرکت رهیاب پیام گستران 2- شرکت رهیاب پیام گستران

کلمات کلیدی :

Machine Learning،Student Enrollment،Explainable AI،Decision Support System،Logistic Regression،XGBoost،Multilayer Perceptron

چکیده :

Artificial Intelligence (AI) has become an essential component in decision-making systems across various domains. In the field of education, explainable machine learning (XML) frameworks enhance transparency and improve data-driven decision-making. An explainable multi-layer architecture is proposed for improving tenth-grade student enrollment decisions using academic and behavioral data from 195 students at Mofid Girls School, Tehran, Iran. The model integrates Decision Tree and Random Forest algorithms for feature selection and applies Logistic Regression, XGBoost, and Multilayer Perceptron (MLP) to predict enrollment outcomes. Among the models, Logistic Regression demonstrated the highest interpretability and accuracy, achieving an accuracy of 0.7599 and an AUC of 0.84. The model was embedded in an interactive chatbot-based decision support system for school administrators.