Stock Market Trend Prediction Using Support Vector Machines and Variable Selection Methods

DOI

10.2991/ammsa-17.2017.45How to use a DOI?

Keywords

machine learning; support vector machines; time series; technical analysis; data mining

Abstract

In this paper, a prediction model integrating machine learning and statistical analysis tools is presented to predict the trend of stock market. The proposed approach consists of three stages, as follows. In the first stage, the system uses technical analysis to calculate useful indicators based on historical data. Then, two different variable selection methods are applied to select the most important variables that describe the given data set. Finally, support vector machines (SVM ) has been used to construct the forecasting model. The hybridized approach was tested to solve the prediction task of directional changes in Dow Jones industrial average (DJIA) index. To evaluate the effectiveness of the use of variable selection techniques in construction of prediction models, this paper compares the performance of the proposed model with the standard SVM-based method. The study concludes that the use of a successful feature extraction technique can improve the forecasting accuracy of the prediction model.

Copyright

© 2017, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Hakob Grigoryan
PY  - 2017/05
DA  - 2017/05
TI  - Stock Market Trend Prediction Using Support Vector Machines and Variable Selection Methods
BT  - Proceedings of the 2017 International Conference on Applied Mathematics, Modelling and Statistics Application (AMMSA 2017)
PB  - Atlantis Press
SP  - 210
EP  - 213
SN  - 1951-6851
UR  - https://doi.org/10.2991/ammsa-17.2017.45
DO  - 10.2991/ammsa-17.2017.45
ID  - Grigoryan2017/05
ER  -