Using wavelets for data generation

BAYAZIT M., Aksoy H.

JOURNAL OF APPLIED STATISTICS, cilt.28, sa.2, ss.157-166, 2001 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 28 Sayı: 2
  • Basım Tarihi: 2001
  • Doi Numarası: 10.1080/02664760020016073
  • Dergi Adı: JOURNAL OF APPLIED STATISTICS
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Sayfa Sayıları: ss.157-166
  • İstanbul Teknik Üniversitesi Adresli: Hayır

Özet

Wavelets are proposed as a non-parametric data generation tool. The idea behind the suggested method is decomposition of data into its details and later reconstruction by summation of the details randomly to generate new data. A Haar wavelet is used because of its simplicity. The method is applied to annual and monthly streamflow series taken from Turkey and USA. It is found to give good results for non-skewed data, as well as in the presence of auto-correlation.