Extraction of relevant dataset for support vector machine using entropy based pivoting QR decomposition and linear discriminant analysis

Support Vector Machines (SVM) are considered to be good machine learning models for classification and regression analysis. However its training time is comparatively higher, especially when the dataset is large. This paper proposes a novel method QRESVM, that reduces the training time, using QR decomposition (QRD) and Linear Discriminant Analysis (LDA) without affecting the classification accuracy. The experimental results indicate that our approach has good classification accuracy with significantly smaller training time.