Package 'TSANN'

Title: Time Series Artificial Neural Network
Description: The best ANN structure for time series data analysis is a demanding need in the present era. This package will find the best-fitted ANN model based on forecasting accuracy. The optimum size of the hidden layers was also determined after determining the number of lags to be included. This package has been developed using the algorithm of Paul and Garai (2021) <doi:10.1007/s00500-021-06087-4>.
Authors: Md Yeasin [aut, cre], Ranjit Kumar Paul [aut], Dipro Sinha [aut]
Maintainer: Md Yeasin <[email protected]>
License: GPL-3
Version: 0.1.0
Built: 2025-03-08 04:01:05 UTC
Source: https://github.com/cran/TSANN

Help Index

Time Series Artificial Neural Network

Description

The best ANN structure for time series data analysis is a demanding need in the present era. This package will find the best-fitted ANN model based on forecasting accuracy. The optimum size of the hidden layers was also determined after determining the number of lags to be included. This package has been developed using the algorithm of Paul and Garai (2021) <doi:10.1007/s00500-021-06087-4>.

Usage

Auto.TSANN(data, min.size, max.size, split.ratio)

Arguments

data

Time Series Data
min.size

Minimum Size of Hidden Layer
max.size

Maximum Size of Hidden Layer
split.ratio

Training and Testing Split Ratio

Value

A list containing:

  • FinalModel: Best ANN model

  • Trace: Matrix of All Iteration

  • FittedValue: Model Fitted Value

  • PredictedValue: Model Forecast Value of Test Data

  • Train.RMSE: Root Mean Square Error of Train Data

  • Test.RMSE: Root Mean Square Error of Test Data

References

Paul, R.K. and Garai, S. (2021). Performance comparison of wavelets-based machine learning technique for forecasting agricultural commodity prices, Soft Computing, 25(20), 12857-12873

Examples

set.seed(16)
x<-rnorm(n = 50, mean = 150, sd = 10)
Auto.TSANN(x,1,2,0.80)