Feature selection for hyperspectral data based on recursive support vector machines

摘要: In this article, a feature selection algorithm for hyperspectral data based on recursive support vector machine (R‐SVM) is proposed. The new follows the scheme of state‐of‐the‐art algorithm, SVM elimination or SVM‐RFE, and uses ranking criterion derived from R‐SVM. Multiple SVMs are used to address multiclass problem. applied Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) select most informative bands resulting subsets compared with SVM‐RFE using accuracy classification as evaluation effectiveness selection. experimental results an agricultural case study indicate that subset generated by newly proposed generally competitive in terms more robust presence noise.