Feature Selection Using Binary Cuckoo Search Algorithm

Kaya Y.

26th IEEE Signal Processing and Communications Applications Conference (SIU), İzmir, Türkiye, 2 - 05 Mayıs 2018 identifier identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Cilt numarası:
  • Doi Numarası: 10.1109/siu.2018.8404843
  • Basıldığı Şehir: İzmir
  • Basıldığı Ülke: Türkiye
  • Karadeniz Teknik Üniversitesi Adresli: Evet

Özet

The cuckoo search algorithm is one of the nature-inspired algorithms developed in recent years. Interesting breeding strategies of cuckoo birds have inspired the algorithm. The binary cuckoo search algorithm is a special version that reduces the search space to binary. In this study, we used binary cuckoo search algorithm in the feature selection process to classify the premature ventricular contraction beats in ECG. We used 200 amplitude values of the signal to represent one beat and selected those that represent the signal best. Using selected features, premature ventricular contraction beats were perceived with 99.67% of accuracy.