Fuzzy Support Vector Regression with Grey Wolf Optimization in Stock Market Prediction

The stock market price is influenced by many factors both domestic and international factors. To help stock traders in make buying or selling decision, a stock price prediction model is needed. In this paper, a stock closing price prediction system is introduced. We utilized 17 financial factors, i.e., 8 technical analysis factors (Crude oil price, Gold spot, Thai Baht to US Dollar exchange rate, High price, Low price, Opening price, Closing price, and Volume) and 9 fundamental factors (Exponential moving average, Relative strength index, Current ratio, Return on equity ratio, Return on assets ratio, Net profit margin ratio, Debt to equity ratio, Price to earning ratio, and Price to book value) from 18 previous working days as our inputs. We then computed chaos centroids from each factor. Finally, to predict the Monday closing price, the fuzzy support vector regression with the grey wolf optimization is utilized. There are 29 companies in 6 industry groups, i.e., RESOURCE, SERVICE, INDUS, PROPCON, AGRO, and TECH, considered in this research. We found that the minimum root mean square error of the blind test set is 0.3698 and the maximum of that is 11.66. However, the prediction trend of each company is very similar to the real closing price.