Comparison of various models on cancer rate and forecasting

Abstract

In this research work, three models were identified; linear regression model, exponential growth model and the quadratic trend model and the results of the work compared. Data collected from Niger State Hospital Management Board was used for the forecast and the result revealed that the quadratic trend model gave the best fit with mean percentage error of -1.08 and mean absolute percentage error of 43.62. Hence, the quadratic trend model was used for the forecast and the result indicated 27.5 percent annual increase in the future trend and the number of patients with cancer that will visit the hospitals within the study area. This figure is well above the national risk of getting cancer before age 75 years fixed at 10.4 percent.

Keywords: Cancer, Tumor, Leukemia, Linear Regression, Mean Percentage Error