Prediction of Stock Index of Tata Steel using Hybrid Machine Learning Based Optimization Techniques
Mohammed Siddique1, Debdulal Panda2
1Mohammed Siddique, Department of Mathematics, Centurion University of Technology and Management, Odisha, India
2Debdulal Panda, Department of Mathematics, KIIT Deemed to be University, Bhubaneswar, India
Manuscript received on 06 March 2019 | Revised Manuscript received on 12 March 2019 | Manuscript published on 30 July 2019 | PP: 3186-3193 | Volume-8 Issue-2, July 2019 | Retrieval Number: B3223078219/19©BEIESP | DOI: 10.35940/ijrte.B3223.078219
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: The trend of stock price prediction has always been in the focal point of analytical activity in financial domain for both the researchers and investors. Prediction with accuracy is very essential for improved investment decisions that imbibe minimum risk factors. Due to this, majority of investors depend upon that intelligent trading system which generates better forecasting results. As forecasting stock market price with high accuracy is quite a challenging task for the analysts, machine learning has been adopted as one of the popular techniques to predict future trends. Even if there are many recognized analytical time series analysis that are categorized either under soft computing or under conventional statistical techniques like fuzzy logic, artificial neural networks and genetic algorithms, researchers have been looking for more appropriate techniques which can exhibit improved results. In this paper, we developed different hybrid machine learning based prediction models and compared their efficiency. Dimension reduction techniques such as orthogonal forward selection (OFS) and kernel principal component analysis (KPCA) are used separately with support vector regression (SVR) and teaching learning based optimization (TLBO) to predict the stock price of Tata Steel. The performance of both the proposed approach is evaluated with 4143days daily transactional data of Tata steels stocks price, which was collected from Bombay Stock Exchange (BSE). We compared the results of both OFS-SVR-TLBO and KPCA-SVR-TLBO hybrid models and concludes that by incorporating KPCA is more practicable and performs better results than OFS.
Index Terms: Forecasting of Stock Market; Orthogonal Forward Selection; Kernel Principal Component Analysis, Support Vector Regression; Teaching–Learning-Based Optimization.
Scope of the Article: Machine Learning