Authors - Ronnel A. dela Cruz Abstract - This study presents a machine learning–based predictive analytics framework[1][2][3] for forecasting faculty promotion outcomes in state universities using institutional performance data and OCR-based document processing[ 4][5]. Faculty demographic information, Individual Performance Commitment Review (IPCR) indicators, and digitized faculty documents were utilized to develop predictive classification models. A dataset consisting of 1,000 faculty records was preprocessed through data cleaning, normalization, feature engineering, and SMOTE balancing applied only to the training dataset. Ada- Boost, Gradient Boosting, and XGBoost classifiers were evaluated using Accuracy, Precision, Recall, F1-score, and ROC-AUC metrics. Among the evaluated models, AdaBoost achieved the strongest performance with 97.33% accuracy and 98.36% ROC-AUC. Feature importance analysis identified teaching effectiveness, curriculum development, and mentorship services as dominant predictors of promotion outcomes. The findings demonstrate the potential of integrating machine learning and OCR-driven document processing to support transparent, scalable, and evidence-based faculty promotion systems in higher education institutions.
Open Zoom