Using Time Series Analysis to Forecast the Number of Registered Cancer Patients in the Municipality of Tobruk

الملخص

This study aims to apply the Box–Jenkins methodology in modeling and analyzing the monthly time series data of cancer patients registered at the Oncology Department of Tobruk Medical Center during the period from January 2018 to December 2024. This time series holds significant importance in supporting healthcare decision-making and strategic planning, particularly in light of the continuous increase in cancer incidence. The analysis was conducted using ARIMA models. A range of candidate models was evaluated based on key statistical accuracy metrics, including the Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), in addition to the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC). Upon comparing the competing models, the results indicated that the most suitable model in terms of both goodness of fit and forecasting accuracy was ARIMA(0,1,1), as it yielded the lowest error metrics and most favorable information criteria values. The predictive power of this model was confirmed through the close alignment between the forecasted and actual values, reflecting its high accuracy in capturing the underlying dynamics of the studied time series. Based on this optimal model, statistical forecasts were generated for the expected number of patients over the future period (2025–2027), providing valuable insight for improving medical planning and healthcare service delivery.