Seasonal time series forecasting by the Walsh-transformation based technique

  • E. BajalinovEmail author
  • Sz. Duleba
Original Article


It is relatively well known that Walsh–Fourier analysis is capable of approximating functions by decomposing them into simple values: 1 and − 1. This method and its valuable characteristics however, are seldom applied in time series forecasting. Moreover, despite the promising applicability of the technique, there is gap in the scientific literature for a comprehensive and detailed introduction and application. Inspired by the results Stoffer (J Time Ser Anal 6:261–267, 1985, J Time Ser Anal 8:449–167, 1987, J Time Ser Anal 11:57–73, 1990, J Am Stat Assoc 86(414):461–479, 1991), Stoffer et al. (J Am Stat Assoc 83:954–963, 1988) and Basu et al. (Electr Power Energy Syst 13(4):193–200, 1991) in the current paper, we aim to describe the theory of Walsh–Fourier analysis, to introduce the so-called Walsh transformation based “row-wise” forecasting process—using the characteristics of the Walsh-matrix—for its application and to compare numerically in R-programming environment (RStudio) using the MComp test data set and the state-of-the-art methods (generic function forecast()) the following two approaches:
  1. (1)

    the 1st of them may be referred to as “direct” since we apply function forecast() to time-series directly in “conventional” mode,

  2. (2)

    and the 2nd one may be called “Walsh-based row-wise” since we apply the same function forecast() to transformed time-series row-wisely.

As can be seen—based on our intentions—a significant advantage of the proposed forecasting process is that when forecasting more than one-step ahead it does not use previously forecasted values and, hence, does not accumulate inaccuracy, thus the forecasted results can be more proper compared to other mainstream techniques. Moreover, as show our numeric results the longer the horizon of forecasting, the more attractive relative accuracy achieved.


Walsh-transform Seasonal time-series analysis Forecasting 


Supplementary material

Supplementary material 1 (R 1 KB)
Supplementary material 2 (R 2 KB)
Supplementary material 3 (R 2 KB)
10100_2019_614_MOESM4_ESM.xlsx (331 kb)
Supplementary material 4 (xlsx 331 KB)
10100_2019_614_MOESM5_ESM.xlsx (442 kb)
Supplementary material 5 (xlsx 441 KB)
Supplementary material 6 (TXT 0 KB)
Supplementary material 7 (TXT 0 KB)


  1. Ahmed N, Schreiber HH, Lopresti PV (1973) On notation and definition of terms related to a class of complete orthogonal functions. IEEE Trans Electromag Cornpat 15:75–80CrossRefGoogle Scholar
  2. Ahmed N, Rao KR (1975) Orthogonal transforms for digital signal processing. Springer, BerlinCrossRefGoogle Scholar
  3. Baghestani H, McNown R (1992) Forecasting the federal budget with time series. J Forecast 11(2):127–139CrossRefGoogle Scholar
  4. Basu TK, Bhattachariya TK, Purkayastha P (1991) Medium range forecasting of hourly power system load by time series analysis using the Walsh transform. Electr Power Energy Syst 13(4):193–200CrossRefGoogle Scholar
  5. Beauchamp KG (1975) Walsh functions and their applications. Academic Press, London, pp 17–20 (Chap. 2)Google Scholar
  6. Beauchamps KG (1984) Applications of Walsh functions and related functions with an introduction to sequency theory. Academic Press, LondonGoogle Scholar
  7. Bichescu BC, Fry MJ, Polak GG (2009) Workload balancing through recurrent subcontracting. Prod Op Manag 18(1):33–47CrossRefGoogle Scholar
  8. Bochner S, Chandrasekharan K (2001) Fourier transforms. Princeton Book Company Publishing, PrincetonGoogle Scholar
  9. Brown RD (1977) A recursive algorithm for sequency-ordered fast Walsh transforms. IEEE Trans Comput C–26(8):819–822CrossRefGoogle Scholar
  10. Broze L, Mélard G (1990) Exponential smoothing: estimation by maximum likelihood. J Forecast 9(5):445–455CrossRefGoogle Scholar
  11. Cryer JD (1986) Time-series analysis Boston. Duxbury Press, DuxburyGoogle Scholar
  12. Falkowski BJ, Sasao T (July 2004) Implementation of Walsh function generator of order 64 using LUT cascades. In: Proceedings of the 47th IEEE international midwest symposium on circuits and systems. Hiroshima, Japan 3:pp 467–470Google Scholar
  13. Falkowski BJ, Sasao T (2005) Unifed algorithm to generate Walsh functions in four different orderings and its programmable hardware implementations. IEE Proc Vis Image Signal Process 152(6):819–826CrossRefGoogle Scholar
  14. Gát Gy (1996) Pointwise convergence of double Walsh-Fejér means. Ann Univ Sci Budapestiensis Sect Comput 16:173–184Google Scholar
  15. Gát Gy, Nagy K (2010) Pointwise convergence of cone-like restricted two-dimensional Fejér means of Walsh-Fourier series. Acta Math Sini Engl Ser 26(12):2295–2304CrossRefGoogle Scholar
  16. Gelper S, Fried R, Croux C (2010) Robust forecasting with exponential and Holt-Winters smoothing. J Forecast 29(3):285–300Google Scholar
  17. Goginava U (2001) On the uniform convergence of Walsh–Fourier series. Acta Math Hunga 93(1–2):59–70CrossRefGoogle Scholar
  18. Golubov B, Efimov A, Skvortsov V (1991) Theory and applications of Walsh series and transforms. Kluwer Academic, BostonCrossRefGoogle Scholar
  19. Harmuth HF (1977) Sequency theory foundations and applications. Academic Press, New YorkGoogle Scholar
  20. Hurst SL, Miller DM, Muzio JC (1985) Spectral techniques in digital logic. Academic Press, LondonCrossRefGoogle Scholar
  21. Hyndman Rob J (2019) Package ‘forecast’—CRAN,
  22. Hyndman Rob J (2018) Forecast package for R,
  23. Hyndman Rob J (2008) Yeasmin Khandakar automatic time series forecasting: the forecast package for R, J Stat Softw 27(3)Google Scholar
  24. Kohn R (1980) On the spectral decomposition of stationary time series using Walsh functions. Adv Appl Probab 12(1):183–199CrossRefGoogle Scholar
  25. Kuklinski WS (1983) Fast Walsh transform data-compression algorithm: E.c.g. applications. Med Biol Eng Comput 21(4):465–472CrossRefGoogle Scholar
  26. Kunt M (1979) In place computation of the Hadamard transform in Cal-Sal order. Signal Process 1(3):227–231CrossRefGoogle Scholar
  27. Kvaratskhelia V (2001) A numerical characteristic of the Sylvester matrix. Discret Math Appl 11(5):501–506CrossRefGoogle Scholar
  28. Kvaratskhelia V (2002) Some inequalities related to Hadamard matrices. Funct Anal Appl 36(1):68–70CrossRefGoogle Scholar
  29. Kvaratskhelia V (2014) Unconditional convergence of functional series in problems of probability theory. J Math Sci 200(2):143–294CrossRefGoogle Scholar
  30. Kvaratskhelia V, Figula A (2015) Some numerical characteristics of Sylvester and Hadamard matrices. Publ Math Debrecen 86(1–2):149–168Google Scholar
  31. Lighthill MJ (2003) Introduction to Fourier analysis and generalised functions. Cambridge University Press, CambridgeGoogle Scholar
  32. Makridakis S, Hibon M (1998) ARMA models and the Box-Jenkins methodology. J Forecast 16(3):147–163CrossRefGoogle Scholar
  33. Manz J (1972) A sequency-ordered fast Walsh transform. Audio Electroacoustics IEEE Trans 20(3):204–205CrossRefGoogle Scholar
  34. Morettin PA (1981) Walsh spectral analysis. SIAM Rev 23(3):279–291CrossRefGoogle Scholar
  35. Morettin PA (1983) A note on a central limit theorem for stationary processes. J Time Ser Anal 4:49–52CrossRefGoogle Scholar
  36. Nakata S (1979) On the unconditional convergence of Walsh series. Anal Math 5:201–205CrossRefGoogle Scholar
  37. Nason GP, Sapatinas T, Sawczenko A (1998) Statistical Modeling of time-series using non-decimated Wavelet representations, CiteSeerX:
  38. Paley REAC (1932) A remarkable series of orthogonal functions (I). Proc London Math Soc 34:241–264CrossRefGoogle Scholar
  39. Shanks JL (1969) Computation of the fast Walsh–Fourier transform. Comput IEEE Trans C–18(5):457–459CrossRefGoogle Scholar
  40. Stankovic RS, Stoic MR, Stankovic MS (1996) Recent developments in abstract harmonic analysis with applications in signal processing. Nauka, BelgradeGoogle Scholar
  41. Stankovic RS, Stankovic M, Jankovic D (1998) Spectral transforms in switching theory, definitions and calculations. Nauka, BelgradeGoogle Scholar
  42. Stoffer DS (1985) Central limit theorems for finite Walsh–Fourier transforms of weekly stationary time series. J Time Ser Anal 6:261–267CrossRefGoogle Scholar
  43. Stoffer DS (1987) Walsh–Fourier analysis of discrete valued time series. J Time Ser Anal 8:449–467CrossRefGoogle Scholar
  44. Stoffer DS, Scher MS, Richardson GA, Day NL, Coble PA (1988) A Walsh-Fourier analysis of the effects of moderate maternal alcohol consumption on neonatal sleep-state cycling. J Am Stat Assoc 83:954–963Google Scholar
  45. Stoffer DS (1990) Multivariate Walsh-Fourier analysis. J Time Ser Anal 11:57–73CrossRefGoogle Scholar
  46. Stoffer DS (1991) Walsh-Fourier analysis and its statistical applications. J Am Stat Assoc 86(414):461–479CrossRefGoogle Scholar
  47. Usha K, Sankar KJ (2012) Generation of Walsh codes in two different orderings using 4-bit gray and inverse gray codes. Indian J Sci Technol 5(3):2341–2345Google Scholar
  48. Walsh JL (1923) A closed set of orthogonal functions. Am J Math 45(1923):5–24CrossRefGoogle Scholar
  49. Wang Z (Aug 1990) In-place and in-order algorithms for Paley- and sequency-ordered Walsh-Hadamard transforms. In: Proceedings of “IEEE international symposium on electromagnetic compatibility”, pp 90,94, 21–23Google Scholar
  50. Yaroslavsky LP (1985) Digital picture processing an introduction. Springer, BerlinCrossRefGoogle Scholar
  51. Yaroslavsky LP (2003) Digital holography and digital image processing: principles, methods, algorithms. Kluwer Academic, BostonGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2019

Authors and Affiliations

  1. 1.University of NyíregyházaNyíregyházaHungary
  2. 2.Budapest University of Technology and EconomicsBudapestHungary

Personalised recommendations