Performance Evaluation of Feature Selection Algorithms on Human Activity Classification


Tulum G., Artug N. T., BOLAT B.

IEEE International Symposium on INnovations in Intelligent SysTems and Applications (INISTA), Bulgaria, 19 - 21 June 2013 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/inista.2013.6577634
  • Country: Bulgaria
  • Yıldız Technical University Affiliated: Yes

Abstract

In this work, four human activities were classified by using multi layer perceptron and k-nearest neighbours algorithm. Due to mass amount of data, two different feature selection methods, which are ReliefF and t-score, were applied to the data. The best result is obtained as 97.6% with 51 features selected by ReliefF.