Stock Price Prediction Using Fuzzy Time-Series Population Based Gravity Search Algorithm

Stock Price Prediction Using Fuzzy Time-Series Population Based Gravity Search Algorithm

Srinivasan N. (Research Scholar, School of Computing, Sathyabama Institute of Science and Technology, Chennai, India) and Lakshmi C. (Prof & Head, Department of Software Engineering, SRM Institute of Science and Technology, Chennai, India)
Copyright: © 2019 |Pages: 15
DOI: 10.4018/IJSI.2019040105

Abstract

The main motive of this research is to predict the future stock value of the particular day with minimum variation from the actual value of stock. In this research, a genetic algorithm-based gravity search algorithm is proposed for stock market prediction. It will be helpful for short-term investors in the National stock market. Some important factors that affect the value of stock are total stocks traded, total turnover of the company, gross domestic product (GDP) of the country, GDP per capita and political or external factors are some of the main factors that affect the stock value of that particular day. Opening and closing values of the stock market were predicted with the help of the above factors. Each factor will be considered as an object with mass, the mass of every object will be based on the importance. With the help of a Gravitational Search Algorithm (GSA) [1], the converging point of the entire object is determined and it is said to be the optimal output of the algorithm. The input considered are opening, closing, low and high values for a period of one year.
Article Preview
Top

Farshchian and Jahan (2015) used Hidden Markov Model (HMM) to predict the changes in Tehran Stock Exchanges. In this technique, the normal factors that affect stock prices are used along with abnormal conditions such as political effect and other factors. All the factors in the data are trained with the help of Baum-Welch Algorithm after that the progressive prediction was achieved by HMM method. The overall accuracy, specificity and sensitivity will be increased up to 2% compared to other previous systems. The periodic accuracy is found to be non-linear compared to original stock value data’s (Saadat & Rahmani, 2016).

Budhani and Budhani (2014) proposed stock prediction method using Artificial Neural Network (ANN) algorithm. ANN is capable of progressive learning so the prediction output will be more efficient than the other soft computing. The main difference between ANN and other soft computing algorithm were the nonlinear behaviour of input dataset with no assumption results. In this technique, the feed forward neural network with back propagation technique was used but the performance of ANN algorithm is not reliable, so it can’t be applied to all the types of input datasets (Budhani, Jha, & Budhani, 2014).

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 8: 4 Issues (2020): 2 Released, 2 Forthcoming
Volume 7: 4 Issues (2019)
Volume 6: 4 Issues (2018)
Volume 5: 4 Issues (2017)
Volume 4: 4 Issues (2016)
Volume 3: 4 Issues (2015)
Volume 2: 4 Issues (2014)
Volume 1: 4 Issues (2013)
View Complete Journal Contents Listing