Prediction of The Topographic Shape of The Ground Surface Using IDW Method through The Rectangular-Neighborhood Approach

The spatial interpolation is to estimate unobserved location variables in geographic space based on the values of the observed location. IDW (Inverse Distance Weighting) is one of the spatial interpolation methods based on the inverse weight of distance between location to be estimated and all observed points. This study applies IDW to predict the topographic shape of the ground surface based on several data samples using the rectangular neighborhood approach. The observation area is a rectangular area generated from several sample points observed which mapped into 50×50 size of X-Y coordinates analogous to an image. This study uses ten samples of observational data from ten different locations. Using only five samples from the existing data samples then there are 2495 new points to be predicted. Logically, of course, there will be improved performance results if using a number of more data samples. Prediction performed by applying the sliding neighborhood operation to observation area using 3×3 rectangular-kernels. For even-numbered data samples as training data have resulted in a predicted success rate of 76.29%, while the odd-numbered data samples have resulted in a predicted success rate of 79.83%.