| Type: | Package |
| Title: | Time Series Intervention Model Using Non-Linear Function |
| Version: | 0.1.0 |
| Author: | Dr. Amrit Kumar Paul [aut], Dr. Md Yeasin [aut, cre], Dr. Ranjit Kumar Paul [aut], Mr. Subhankar Biswas [aut], Dr. HS Roy [aut], Dr. Prakash Kumar [aut] |
| Maintainer: | Dr. Md Yeasin <yeasin.iasri@gmail.com> |
| Description: | Intervention analysis is used to investigate structural changes in data resulting from external events. Traditional time series intervention models, viz. Autoregressive Integrated Moving Average model with exogeneous variables (ARIMA-X) and Artificial Neural Networks with exogeneous variables (ANN-X), rely on linear intervention functions such as step or ramp functions, or their combinations. In this package, the Gompertz, Logistic, Monomolecular, Richard and Hoerl function have been used as non-linear intervention function. The equation of the above models are represented as: Gompertz: A * exp(-B * exp(-k * t)); Logistic: K / (1 + ((K - N0) / N0) * exp(-r * t)); Monomolecular: A * exp(-k * t); Richard: A + (K - A) / (1 + exp(-B * (C - t)))^(1/beta) and Hoerl: a*(b^t)*(t^c).This package introduced algorithm for time series intervention analysis employing ARIMA and ANN models with a non-linear intervention function. This package has been developed using algorithm of Yeasin et al. <doi:10.1016/j.hazadv.2023.100325> and Paul and Yeasin <doi:10.1371/journal.pone.0272999>. |
| License: | GPL-3 |
| Encoding: | UTF-8 |
| Imports: | stats, forecast, MLmetrics |
| RoxygenNote: | 7.2.1 |
| NeedsCompilation: | no |
| Packaged: | 2024-04-18 09:01:40 UTC; YEASIN |
| Repository: | CRAN |
| Date/Publication: | 2024-04-18 19:13:03 UTC |
Time Series Intervention Model Using Non-linear Function
Description
Time Series Intervention Model Using Non-linear Function
Usage
InterNL(Data, Time, TSModel, TSOrder = NULL, NLModel, InitialNLM)
Arguments
Data |
Time series data |
Time |
Point of intervention |
TSModel |
Time series model ("arima" or "ann") |
TSOrder |
If model is ANN, then order is lag of the model |
NLModel |
Non-linear models ("gompertz","logistic", "monomolecular", "richard", "hoerl") |
InitialNLM |
Initial value for parameters of non-linear model |
Value
Accuracy: Accuracy metric of the proposed model
PreFitted: Fitted values for the pre intervention series
PostFitted: Prediction for the post intervention series
NLM: Details of fitted non-linear model
References
Paul, R.K. and Yeasin, M., 2022. COVID-19 and prices of pulses in Major markets of India: Impact of nationwide lockdown. Plos one, 17(8), p.e0272999.
Yeasin, M., Paul, R.K., Das, S., Deka, D. and Karak, T., 2023. Change in the air due to the coronavirus outbreak in four major cities of India: What do the statistics say?. Journal of Hazardous Materials Advances, 10, p.100325.
Examples
library("InterNL")
data<- as.ts(rnorm(120,100,50))
Result <- InterNL(Data = data,Time = 90, TSModel = "arima",
TSOrder=NULL, NLModel=NULL, InitialNLM=NULL )