| Average and exponentially weighted return ratios are studied to forecast undergraduate student enrollment in a university. The objective is to develop a low-cost and user-friendly forecasting model to reduce forecasting error. Based on analysis of different forecasting techniques, three forecasting models (i.e., university-level analysis, school-level analysis, and division-level analysis) are compared using average and exponentially weighted return ratios. Different years of average return ratios have also been discussed to compare the forecasting models. The experimental results indicate that when forecasting for spring semesters, one year average return ratio using school-level analysis should be applied. When forecasting for fall semesters, average return ratio method using university-level analysis is recommended. The results also show that forecasting error is not reduced when using division-level analysis versus school-level analysis and university-level analysis. It is also noticed that forecasting error is not reduced when using all available enrollment data versus the most recent enrollment data. |
