Analysis of lignite quality variables via geographical weighting

In this study, forcefulness of geographically weighting in heterogeneous mine deposits has been investigated. The relationships among the lignite quality variables such as moisture, ash, volatile matter, sulphur content and calorific values are appraised by spatial information. By this way, it is aimed to incorporate geological uncertainties in model processes. In the first part of the thesis, to identify the relationships at local and global scale, in addition to the deterministic methods (ISDW), geostatistical approach (Kriging) and regression-based weighting models (GWR, Local Regression) have been employed. In the second stage, an alternative to conventional linear model solution, the generalized additive model approach (GAM) has been used and the effectiveness of spatial relationships in the lignite site has been investigated. The findings acquired have also enabled a possibility to compare the models along with the evaluation of parameter efficiency in the lignite site. Geographically weighting is an effective tool for determining the lignite zone of influences in accordance with the variable combinations and improving the linear model performances.