The Application of Rough Set Theory in Stock Price Forecasting (Case Study: Iran Saderat Bank)

Abstract

This paper proposes a method based on rough set theory and by using technical indicators to predict the stock price. Rough set theory has several advantages; the most important advantage is that no additional information is needed about the initial data. In the proposed model, a number of technical indicators from the data of Bank Saderat Iran during a year were calculated and used as condition attributes in the decision table and the stock price fluctuation on the next day was selected as decision attribute. It should be noted that by using the correlation matrix analysis, the variables with the highest correlation with decision attribute were selected as conditional attributes. Using rough set theory and different discretization and reduction methods, some rules are extracted based on learning data and methods validity were computed based on control data. Comparing the return of this method and buy and hold method reveals the superiority of proposed model. Also, using data from different years with different price trends as inputs to the model and achieving satisfactory results is a promising reason for using and developing this method in stock price forecasting.

Keywords

• Stock price forecasting
• Rough set theory
• Conditional and Decision Attributes
• Data Discretization
• Rules Extraction