TESTING SIMPLE REGRESSION MODEL FOR COORDINATE TRANFORMATION BY COMPARING ITS PREDICTIVE RESULT FOR TWO REGIONS

Coordinate Transformation between different datum usually involve complicated equation and computational analysis that is time consuming. Moreover, systematic errors are inevitable in the course of the transformation process thereby leading to accuracy problem. The theoretical mathematical models are methods most surveyors are versatile with. As more non professional surveying discipline is being introduced to GPS/GIS technology, the need for simple coordinate transformation from global WGS84 to local planimetric coordinates is absolutely important to guarantee local positioning accuracy. The research has been able to compare the validity of regression models developed from random distribution of GPS points in two regions. The points distribution in the first region are relatively close with an average distance less than 350m while for the second region are relatively far apart with an average distance of at least 850m. The model predictive values from the region, whose random point distributions are close, are far better with minimal errors compared with theoretical model. While the result from region whose GPS points are far apart was unsatisfactory, pointing out factor to be considered when developing a regression model for coordinate transformation.