Optimization of Markov chain modeling in predicting college student retention

College student retention is one of the most important metrics in higher education. With institutions across the US facing decreasing enrollment, developing a reliable retention prediction method is crucial. In recent years, the use of the Markov chain model in forecasting student enrollment and progression has become more common, but there is little work on its application in student retention. One key factor in determining this model's effectiveness is what parameters should be used in the student population's segmentation or grouping. This study presents a rigorous algorithm, coupled with a prediction model, capable of selecting parameters that provide the most accurate results for term-to-term retention prediction using the Markov chain analysis. This is a pioneering attempt to optimize the Markov chain for retention prediction. Results are verified against an enrollment dataset from a public institution. With high accuracy, flexibility, and interpretability, the coupled algorithm model is an effective tool for institutional planning. Future iterations will focus on adding behavioral data and integrating the model with advanced deep learning methods to predict the detailed status of returning students, such as full-time equivalent and class level.