Modeling Breast Cancer Cases in Fiji

Breast cancer incidences is steadily increasing in Fiji and accurate forecasting can have major implications in controlling this deadly illness. The best forecasting model is the one that does not underestimates or overestimates the true number of breast cancer cases and gives minimal prediction errors. This paper proposes Linear Regression model for forecasting breast cancer cases using the reported number of cases in Fijian population from the year 1995 to 2016. The proposed model is also compared with the Naïve Forecast Method as the benchmark. The performances of the two models were analyzed based on measures such as the Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE). The proposed model was then further validated using the diagnostic measures such as Goodness-of-fit (R²), Tracking Signal (TS) and Bias. The results showed that the proposed Linear Regression model outperformed the Naïve Forecast Method. It also satisfies the validity diagnostic measures and is a better tool in forecasting Fiji's yearly breast cancer cases.