Satellite Data Classification Based On Support Vector Machine, Rough Sets Theory & Rough-SVM

Abstract

As Classification is becoming one of the most crucial tasks for various applications. Text categorization, tone recognition, image classification, micro-array gene expression are the few examples of such kind of applications. The supervised classification techniques are mostly based on traditional statistics capable of providing good results when sample size seems to tend to infinity. But in practice, only finite samples can be acquired. In this paper, an innovative learning technique, Rough Support Vector Machine (SVM), is employed on Satellite Data multi class. SVM Initiated in the early 90?s, a powerful machine technique amplified from arithmetical learning led to an outburst of interest in machine learning and have made noteworthy achievement in some field as SVM technique does not agonize the boundaries of data dimensionality and limited samples [1] & [2]. In our investigation, as the support vectors, classification are gathered by learning from the training samples are very perilous. In this paper, using various kernel functions for satellite data samples relative outcomes explained.