Batra, LuckshayTaneje, H.C.2024-07-122024-07-122019Batra, L. ve Taneja, H. C. (2019). Comparison of ARIMA, holt-winters, and LSTM forecasting models using kullback information measure. International Conference of Mathematical Sciences (ICMS 2019). s. 189.978-605-2124-29-1https://hdl.handle.net/20.500.12415/2091Various forecasting models such as the Autoregressive Integrated Moving Average (ARIMA) and HoltWinters that aren’t just widely accepted but also exceptionally good predictors of the time series. Recently, Artificial Neural Networks (ANNs) have been widely studied and utilized in the prediction of time series, and their flexible non-linear modeling capacity is the key advantage of deep learning. Long Short Term Memory (LSTM), in particular has been used in the prediction of time-series in financial sector. The objective of this study is to examine and compare different forecasting models in terms of performance on a time series that is considered difficult to predict. This article’s core contribution is to contrast the performance of ARIMA, Holt-Winters and a recurrent neural network LSTM with reference to minimization obtained in the Kullback measure of relative information in prediction. The results shows that LSTM network performs well on monthly data from the NIFTY 50 stock index, a real-life time series forecast in comparison with traditional models like ARIMA and Holt-winters.enCC0 1.0 Universalinfo:eu-repo/semantics/openAccessShannon entropyKullback measure of relative informationTime series forecastingComparison of ARIMA, holt-winters, and LSTM forecasting models using kullback information measureArticle190189