Statistical Evaluation and Prediction of Financial Time Series Using Hybrid Regression Prediction Models
DOI:
https://doi.org/10.18201/ijisae.2021473645Keywords:
LSTM, Chaos, Polynomial Regression, Financial Time Series (FTS), Deep Learning, Time Series Prediction, Exchange rate, Stock market index, Commodity price.Abstract
Financial time series (FTS) is chaotic which contributes to dynamic and difficult predictability in turn. This research employs the new hybrid model for the prediction of FTS which comprises Long Short-Term Memory (LSTM), Polynomial Regression (PR), and Chaos Theory. First of all, FTS is tested in this hybrid for chaos. Later, using Chaos Theory, the time series is modelled with the chaos existence. The model time series will be entered in LSTM for initial forecasts. The sequence of errors derived from LSTM forecasts is PR appropriate for error predictions. Error forecasts and original model forecasts are applied to produce the final hybrid model forecasts. Performance testing of the hybrid model (Chaos+LSTM+PR) is conducted using three categories of foreign exchange, commodity price and stock-market indices. The hybrid model proposed in this study, in compliance with MSE, Dstat and Theil’s U is proved superior to the individual models like ARIMA, Prophet, LSTM and Chaos+LSTM. However, this study of the hybrid method in the stock market indices of Nifty 50 result is equated with the Covid pandemic results using descriptive statistics and also compared with the stock price of Nifty 50 using random forest method from which the result of proposed method gets better prediction than the existing methods from the comparison. The execution of these various hybrid proposed methods is done using mainly in Python, additionally, the authors used Gretl® and R for some methods respectively. Ultimately, the final result of this hybrid model describes with a better result than the existing prediction models and it is proved using various types of FTS like Foreign exchange rates, commodity prices, and stock market indices respectively. Hence, the Diebold Mariano test is carried out to compare the existing and proposed models, certainly that the Chaos+LSTM+PR is better than the existing models on the nine datasets, respectively.
Downloads
Downloads
Published
How to Cite
Issue
Section
License
All papers should be submitted electronically. All submitted manuscripts must be original work that is not under submission at another journal or under consideration for publication in another form, such as a monograph or chapter of a book. Authors of submitted papers are obligated not to submit their paper for publication elsewhere until an editorial decision is rendered on their submission. Further, authors of accepted papers are prohibited from publishing the results in other publications that appear before the paper is published in the Journal unless they receive approval for doing so from the Editor-In-Chief.
IJISAE open access articles are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. This license lets the audience to give appropriate credit, provide a link to the license, and indicate if changes were made and if they remix, transform, or build upon the material, they must distribute contributions under the same license as the original.