Title:	Support vector machines for pattern recognition
DNr:	SNIC 2019/7-76
Project Type:	SNIC Small Compute
Principal Investigator:	Mohammad Aslani <Mohammad.Aslani@hig.se>
Affiliation:	Högskolan i Gävle
Duration:	2019-10-07 – 2020-11-01
Classification:	10201
Keywords:

Abstract

Support vector machine is one the best classifiers in the world of machine learning. However, its time complexity O(n^3) hinders obtaining the results when facing with huge datasets. One of the solutions is to select the instances (instance selection) that play a significant role in classification and remove the rest. In this project, we develop a novel and fast instance selection method. [Abstract has been edited after approval (reference number 199743). Contact support@supr.snic.se if you have questions.]