Regression Models for 2-Dimensional Cartesian Coordinates Prediction: A Case study at University of Mines and Technology (UMaT), Tarkwa-Ghana.

Abstract

The aim of this research is to study and analyze statistical models applicable in bringing out a relationship between global coordinates and cartesian planimetric coordinates of some known control stations in the University of Mines and Technology (UMaT) campus. To achieve the aims of this research, the Global Position System (GPS) latitudes and longitudes of selected control stations with known cartesian planimetric coordinates were determined using the Handheld GPS receiver at different epoch (morning and evening). Linear Regression analysis was then conducted to establish the correlation between global and cartesian planimetric coordinates of the selected control stations and regression models generated to show the results. The correlation coefficient r, a t-test for non -zero slope, t-test on correlation coefficient, graphical residual analysis, test of normality, comparing model predictions to observed data, were used to evaluate and check the adequacy of the models. The obtained results indicated that the proposed linear regression models are suitable for predictions at 95% confidence interval and do not violate any of the statistical assumptions of a linear model. However, the proposed regression models for the evening observation gave better prediction accuracy than the morning. A computer programming algorithm and a designed interface was created for the proposed regression models established using Microsoft C++ standard edition 6.0, thus making it easier in applying the models in making cartesian planimetric coordinates prediction at different epoch at UMaT.