Package 'WaveletRF'

Title: Wavelet-RF Hybrid Model for Time Series Forecasting
Description: The Wavelet Decomposition followed by Random Forest Regression (RF) models have been applied for time series forecasting. The maximum overlap discrete wavelet transform (MODWT) algorithm was chosen as it works for any length of the series. The series is first divided into training and testing sets. In each of the wavelet decomposed series, the supervised machine learning approach namely random forest was employed to train the model. This package also provides accuracy metrics in the form of Root Mean Square Error (RMSE) and Mean Absolute Prediction Error (MAPE). This package is based on the algorithm of Ding et al. (2021) <DOI: 10.1007/s11356-020-12298-3>.
Authors: Ranjit Kumar Paul [aut, cre], Md Yeasin [aut]
Maintainer: Ranjit Kumar Paul <[email protected]>
License: GPL-3
Version: 0.1.0
Built: 2024-10-31 21:07:45 UTC
Source: https://github.com/cran/WaveletRF

Help Index

Wavelet Transform Using Maximal Overlap Discrete Wavelet Transform (MODWT) Algorithm

Description

Transforms the time series data by using hybrid MODWT algorithm.

Usage

WaveletFitting(
  ts,
  Wvlevels,
  WFilter = "haar",
  bndry = "periodic",
  FFlag = TRUE
)

Arguments

ts

Univariate time series
Wvlevels

The level of wavelet decomposition
WFilter

Wavelet filter use in the decomposition
bndry

The boundary condition of wavelet decomposition:'periodic' or 'reflection'
FFlag

The FastFlag condition of wavelet decomposition: True or False

Value

  • WaveletSeries - The wavelet trasnform of the series

References

  • Aminghafari, M. and Poggi, J.M. 2007. Forecasting time series using wavelets. Internationa Journal of Wavelets, Multiresolution and Inforamtion Processing, 5, 709 to 724

  • Percival D. B. and Walden A. T. 2000. Wavelet Methods for Time-Series Analysis. Cambridge Univ. Press, U.K.

  • Paul R. K., Prajneshu and Ghosh H. 2013. Wavelet Frequency Domain Approach for Modelling and Forecasting of Indian Monsoon Rainfall Time-Series Data. Journal of the Indian society of agricultural statistics, 67, 319 to 327.

Examples

data<-rnorm(200,mean=20,sd=3)
Wavelet<-WaveletFitting(ts=data,Wvlevels=3,WFilter='haar',bndry='periodic',FFlag=TRUE)

Wavelet-RF Hybrid Model for Forecasting

Description

The Wavelet Decomposition followed by Random Forest Regression (RF) models have been applied for time series forecasting. The maximum overlap discrete wavelet transform (MODWT) algorithm was chosen as it works for any length of the series. The series is first divided into training and testing sets. In each of the wavelet decomposed series, the supervised machine learning approach namely random forest was employed to train the model. This package also provides accuracy metrics in the form of Root Mean Square Error (RMSE) and Mean Absolute Prediction Error (MAPE).

Usage

WaveletFittingRF(
  ts,
  tlag = ACF,
  Waveletlevels,
  WaveletFilter = "haar",
  boundary = "periodic",
  FastFlag = TRUE,
  SplitRatio = 0.8
)

Arguments

ts

Univariate time series
tlag

Number of lags
Waveletlevels

The level of wavelet decomposition
WaveletFilter

Wavelet filter use in the decomposition
boundary

The boundary condition of wavelet decomposition
FastFlag

The FastFlag condition of wavelet decomposition: True or False
SplitRatio

Training and testing data split

Value

  • TrainFittedValue - Fitted value of train data

  • TestPredictedValue - Predicted value of test data

  • AccuracyTable - RMSE and MAPE of train and test data

References

  • Ding, Y., Zhang, W., Zhao, X., Zhang, L. and Yan, F., 2020. A hybrid random forest method fusing wavelet transform and variable importance for the quantitative analysis of K in potassic salt ore using laser-induced breakdown spectroscopy. Journal of Analytical Atomic Spectrometry, 35(6), 1131-1138.

  • Rezaali, M., Fouladi-Fard, R., Mojarad, H., Sorooshian, A., Mahdinia, M. and Mirzaei, N., 2021. A wavelet-based random forest approach for indoor BTEX spatiotemporal modeling and health risk assessment. Environmental Science and Pollution Research, 28(18), 22522-22535.

  • Paul, R. K., Prajneshu and Ghosh H. 2013. Wavelet Frequency Domain Approach for Modelling and Forecasting of Indian Monsoon Rainfall Time-Series Data. Journal of the Indian society of agricultural statistics, 67, 319-327.

  • Paul, R. K. and Birthal, P.S. 2015. Investigating rainfall trend over India using wavelet technique. Journal of Water and Climate Change, 7, 365 to 378.

  • Paul, R. K. 2015. ARIMAX-GARCH-WAVELET Model for forecasting volatile data. Model Assisted Statistics and Application, 10, 243 to 252.

Examples

data<-rnorm(200,mean=20,sd=3)
WRF<-WaveletFittingRF(ts=data,tlag=2,Waveletlevels=3)