MODIFICATION OF ESTIMATION METHODS FOR INVERSE DISTANCE SQUARE AND NEAREST NEIGHBORHOOD POINT USING BLOCK DISCRETIZATION

• Main Author: Alim, Muhammad
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/54432
• Institution: Institut Teknologi Bandung

Inverse Distance Square (IDS) is an interpolation using the Inverse Distance Weighting (IDW) method which has the value of the Power Function equal to 2 or a power of 2 (squared) while Nearest Neighbor Point is an estimation method where the estimation value is taken based on the effect of each point follow the closest point. IDS and NNP can be discretized by dividing the space into a series of points to turn a coarse grid into a corresponding set of fine grids. Each discretized point or node represents a volume that is an integral part of a grid block. In this study, to facilitate the concept of data processing, artificial data on coordinate of DH1 levels (25; 130) 2.00%; DH2 (87; 125) 3.00%; DH3 (35:28) 3.50%; DH4 (86; 30) 2.50% G1 (50; 70) 2.85%; G2 (70; 90) 2.80% with G1 and G2 as points to be estimated in the experiment. The estimated values of G1 and G2 will then be compared with the actual values using cross validation root mean square error (RMSE). Using the same concept, a number of coal data sets will be processed which have the physical properties of thickness (m), ash content (% w), sulfur content (% w) and calorific values (kcal) of the coal. The data that has been processed is then analyzed. The results showed that the increase in the amount of data that affected the calculation at the time of estimation resulted in the resulting estimated value following the actual data trend and making the data tighter. The block estimation grid using the Invers Distance Square method in this experiment, can represent the sulfur value at the midpoint of the estimation block.