Affiliations: [a] Department of Information Technology, School of Computing and IT, Manipal University Jaipur, Dehmi Kalan, Near GVK Toll Plaza, Jaipur, Rajasthan, India
| [b] Department of Computer Science and Engineering, School of Computing and IT, Manipal University Jaipur, Dehmi Kalan, Near GVK Toll Plaza, Jaipur, Rajasthan, India
Correspondence:
[*]
Corresponding author. Shikha Mundra, Department of Computer Science and Engineering, School of Computing and IT, Manipal University Jaipur, Dehmi Kalan, Near GVK Toll Plaza, Jaipur, Rajasthan, India. E-mail: [email protected].
Abstract: Stock market analysis or stock price prediction is aimed at predicting firm’s profitability based on current as well as historical data. From recent studies it is observed that machine learning approaches have outperformed traditional statistical methods in predictive analysis task. In our work we have analyzed time series data as prediction of stock price depends on historical variation in prices of stocks. To enhance the prediction accuracy, we have proposed a hybrid approach which is based on the concept of support vector machines (SVM) and Long Short-Term Memory (LSTM) as these algorithms are performing better in time series problem. On applying proposed approach onto the TATA Global Beverages stock dataset, we have observed prediction accuracy of ninety seven percent which is outperforming, along with this to enhance the performance author have presented some observation like relative importance of the input financial variables and differences of determining factors in market comparative predictive analysis onto the experimentation dataset.
Keywords: SVM, LSTM, back propagation, RNN, machine learning
