Current Version: 5.1 Date: 2021-01
Authors: Rebecca Pontes Salles ([email protected]) and Eduardo Ogasawara ([email protected])
Description: Functions for defining and conducting a time series prediction process including pre(post)processing, decomposition, modeling, prediction, and accuracy assessment. The generated models and their yielded prediction errors can be used for benchmarking other time series prediction methods and for creating a demand for the refinement of such methods. For this purpose, benchmark data from prediction competitions may be used.
Available at CRAN: https://CRAN.R-project.org/package=TSPred
Reference manual: TSPred.pdf
Acknowledgements: The authors thank CNPq, CAPES, and FAPERJ for partially sponsoring this research.
#Install TSPred package
> install.packages("TSPred")
#Load TSPred package
> library("TSPred")#loading CATS dataset
> data("CATS")
#defining the time series application
> tspred_arima <- tspred( subsetting = subsetting(test_len = 20),
modeling = ARIMA(),
evaluating = list(MSE = MSE(),AIC = AIC()) )
#performing the prediction application and obtaining results
> tspred_arima_res <- workflow( tspred_arima, data = CATS[5] )#Obtaining objects of the processing class
> proc_subset <- subsetting( test_len = 20 )
> proc_bct <- BCT()
> proc_wt <- WT( level = 1, filter = "bl14" )
> proc_sw <- SW( window_len = 6 )
> proc_mm <- MinMax()
#Obtaining objects of the modeling class
> modl_nnet <- NNET( size = 5, sw = proc_sw, proc = list(MM = proc_mm) )
#Obtaining objects of the evaluating class
> eval_mse <- MSE()#Defining a time series prediction process
> tspred_mlm <- tspred( subsetting = proc_subset,
processing = list(BCT = proc_bct, WT = proc_wt),
modeling = modl_nnet,
evaluating = list(MSE = eval_mse) )
#Running the time series prediction process and obtaining results
> tspred_mlm_res <- tspred_mlm %>%
subset(data = CATS[5]) %>%
preprocess(prep_test = TRUE) %>%
train() %>%
predict(input_test_data = TRUE) %>%
postprocess() %>%
evaluate()
#Benchmarking tspred objects
> bmrk_results <- benchmark( tspred_arima_res, list(tspred_mlm_res) )#Subclass my.model
> my.model <- function(train_par=NULL, pred_par=NULL){
MLM(train_func = my.model.func, train_par = c(train_par),
pred_func = my.model.pred.func, pred_par = c(pred_par),
method = "Name of my model", subclass = "my.model" )
}
#Obtaining an instance of the subclass my.model
> model <- my.model(train_par = list(par1="a", par2="b"), pred_par = list(par3="c"))- LogT - Logarithmic transform.
- BCT - Box-Cox transform.
- an - Adaptive normalization.
- Diff - Differencing. MAS
- mas - Moving average smoothing.
- pct - Percentage change transform.
- WaveletT - Wavelet transform.
- emd - Empirical mode decomposition.
- fittestLM - Automatically finding fittest linear model for prediction.
- fittestArima - Automatic ARIMA fitting, prediction and accuracy evaluation.
- fittestArimaKF - Automatic ARIMA fitting and prediction with Kalman filter.
- fittestPolyR - Automatic fitting and prediction of polynomial regression.
- fittestPolyRKF - Automatic fitting and prediction of polynomial regression with Kalman filter.
- fittestMAS - Automatic prediction with moving average smoothing.
- fittestWavelet - Automatic prediction with wavelet transform.
- fittestEMD - Automatic prediction with empirical mode decomposition.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent