A Spectrophotometric Evaluation of Lunar Catharina Crater Using Support Vector Regression Analysis for FeO and TiO₂ Estimations

Abstract

Support Vector Regression (SVR) is an extended version of the Support Vector Machine (SVM) algorithm. It is an effective machine learning tool for handling huge complex data sets. SVR algorithm is introduced to the existing lunar FeO and TiO2 concentrate estimation techniques. This machine learning algorithm is capable of transforming complex nonlinear problems into a higher dimensional feature space and solving it linearly. The SVR analysis of Moon Mineralogy Mapper (M3) data for lunar mineral concentrate estimation shows an upgraded result over the existing estimation methods. Outlier points are less sensitive to SVR and, hence it provides the best fit line or curve.