<![CDATA[New student admission is an important aspect in higher education management, including Nahdlatul Ulama University (UNU) Pasuruan. Relevant prediction of total new students is needed to support resource planning such as teaching staff, facilities, and budget. This study aims to evaluate the historical pattern of new student admissions at UNU Pasuruan and predict the number of new students in the coming years using the ARIMA (Auto Regressive Integrated Moving Average) method. The data used is historical data on new student admissions in the last five years, which is analyzed to identify trends, seasonality, and fluctuation patterns. The analysis is performed using statistical software such as Python to improve the accuracy and efficiency of the process. This study approach includes several main steps, namely collecting historical data on the number of new students, testing stationarity using the Augmented Dickey-Fuller (ADF) test, identifying model parameters through ACF and PACF graphs, and estimating ARIMA model parameters. The resulting model is evaluated using prediction error metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE). The study findings describe that the ARIMA model (6,0,1) produces an RMSE value of 21.88 and a MAPE of 0.2%. In addition to having the smallest error score, the ARIMA model (6,0,1) also has the smallest AIC score of the various models that can be used for predictions, which is 447.44 and the largest log likelihood value, which is -214.72. The largest prediction of the number of new students is in July, which is 92.72 and the smallest in February, which is 24.43. This prediction is expected to help university management in optimizing resource planning, increasing management efficiency, and anticipating fluctuations in the number of new students in the future. This study offers new findings in the form of the use of predictive models based on historical data to support strategic decision- making, such as resource allocation and promotion planning. With these results, universities can anticipate changes in the number of enrollments more effectively, which were previously only done based on subjective estimates. The model built can also be applied to similar datasets in the future with appropriate parameter adjustments.]]>
