Akademik Veri Yönetim Sistemi
18th International Multidisciplinary Scientific Geoconference, SGEM 2018, Albena, Bulgaristan, 2 - 08 Temmuz 2018, cilt.18, ss.551-556
© SGEM 2018.The purpose of this study is to compare different classification techniques which are Maximum Likelihood Classification (MLC), Support Vector Machine (SVM) and Random Forest (RF) for lithological classifications in the western part of Antalya province. During the study, main data source was used as Terra/ASTER image data. The study area (Western Antalya) contains eight different lithological units (i.e. limestone, travertine, peridotite, mélange, sandstone, chert, clastic, and alluvium) which are identified in 1:25 000 scaled geological map and observations from the field study. As a result of the processes, there classifications were produced. When comparing to the performances of the obtained results, MLC and RF are close to each other in terms of their overall accuracy values as 81.72 percent and 81.98 percent, respectively. The SVM technique is more accurate than the other techniques. It has value of 84.64 percent. The most successful classified lithological unit is peridotite for all classifiers used in the study. The classification accuracies of the peridotite unit are obtained as 98.19 percent for MLC, 98.41 percent for SVM, and 98.41 percent for RF. Melange unit is the worst classified lithological unit.