Performance Comparison of FPGA based Linear SVMS Classifier and Computer Simulation

Machine learning algorithms are the turf of artificial intelligence which handle the learning aspect of computers/machines due to advancement in technologies that permitted these binary machines to build a better understanding of patterns and logic; also help in finding solutions for real-world problems. As a machine learning tool, support vector machines (SVMs) are a prominent supervised algorithm that deals with the classification and regression of the observed datasets. In this paper, support Vector Machine algorithm has been implemented over Field Programmable Gate Array (FPGA) for classification in linear mode to accelerate the computations by the aid of the parallel nature of FPGAs to acquire high prediction accuracy at a cost of its high computational complexity.