| Abstrakt: |
Investment in the stocks exchange is one of the most common attitudes in the capital market. If one has the knowledge and the experience, he has the chance to make a lot of profits from this risky market. On the other hand, since stock market has a nonlinear and chaotic behavior, one mistake is enough to lose. This dynamic market is greatly affected by psychological, economic, and political conditions. Scientific methods and practical techniques may be used to carefully investigate the issue. To mention, nonlinear intelligent methods such as metaheuristics, learning automata, and artificial neural network are some of well-known methods which can be used to predict share price before the actual investment. In order to realize that, a model for predicting the share price has been developed and presented by combining the aforementioned methods. Firefly Algorithm has been taken into account as the metaheuristic solution to the stock prediction problem. Such a hybrid method has been proposed and implemented; in order to evaluate the proposed solution, a dataset gathered from 20 companies listed on Tehran Stock Exchange has been considered as a case study. It has been shown that the proposed algorithm can approximately predict 30 to 50% of the future behavior of shares in the market. The outcomes have shown great potential for the hybrid algorithm to be applied in other fields of business administration and investment problems. [ABSTRACT FROM AUTHOR] |
