The application of wavelet transform to analyze the rainfall data

Abstract

We transformed the rainfall data by Haar and Daubechies wavelet func- tion. Decompose Haar and Daubechies to scale and translate for constructing an orthogonal basis and also estimated a function to consist with the real value. Then, continue with ARIMA model to approximate and compare with the minimum value of mean ab solute error (MAE)and root mean error (RMSE) to make a forecast in the future. We can see that the fitted ARIMA model of Haar and Daubechies discrete wavelet transformed data gives the smaller value of mean absolute error (MAE) and root mean error (RMSE) more than ARIMA of rainfall data. However, the model of Daubechies wavelet trans- formed data and gives better result more than Haar wavelet transformed data.