Destek Vektör Makineleri Üzerine Bir Çalışma

Abstract

Support Vector Machines were introduced by Vladamir Vapnik in 1995 as a new-generation technique, learning from training set which is high-dimensional and whose sample size is much small, by creating and solving a quadratic programming problem using some novel mathematical optimization techniques. Statistical Learning Theory, which Support Vector Machines algorithm is based on, has been proposed in 1960s by Vladamir Vapnik and Alexey Chervonenkis and has been literally developed in 1970s. Due to the use of kernel functions to model non-linearity, high performance of generalization, powerful theoretical foundations and the ability to train relatively quickly, in the last decades, this method has been utilized frequently and mostly on pattern recognition, regression analysis, face recognition, image and text classification, data mining, quality control and applications of finance, economy, genetic, biology and bioinformatic.