Forecasting Time Series Data Using Haar Discrete Wavelet Transformation

Discrete Wavelet Transform is a data transformation method that represents in the time domain and frequency domain. This appears to overcome weakness of Fourier transform which only able provide one information limited certain windowing . The type wavelet used Haar Wavelet. Identification periodicity using Periodogram analysis with Fisher's Test statistics. transformed decomposed into two components, namely Approximation Coefficient Detail Coefficient. Both components are predicted Box-Jenkins ARIMA method. Model selection was carried out Akaike Information Criterion (AIC ) Mean Square Error (MSE) methods forecast obtained then reconstructed (inverse). application model through Makassar City Air Humidity for period September 2006 - December 2012 shows forecasting on by (0,0,3) AIC = 112.2142 MSE 29.673. While Detailed Coefficients (2,1,0) 89.2 15,989.