Stock Price Forecasting: Hybrid Model of Artificial Intelligent Methods
Keywords:stock price, Hybrid model, artificial intelligence, multi-agent simulation, Engineering Economics
Predicting the price of stock is very helpful and can attract the interest of researchers and investors who make subjective investment judgments based on objective technical indicators. We propose a new hybrid forecasting model utilizing the combined prediction’s principle as well as the artificial intelligence’s technique. First, the new method presented combines the single models in series. Next, we present the principle of hybrid prediction which is a foundation of our model, and its validity is proven and carefully illustrated in this paper. According to this principle, the combined model consists of three sole prediction models and demonstrates greater forecast accuracy than the single model. The new model also owns the qualitative prediction, as well as the quantitative prediction. Furthermore, a clear introduction of the three sole prediction models is given, and the comprehensive forecasting rules are introduced. The comparison analysis is included in this paper to validate the new method, and the multi-agent simulation, including different investment strategies, is given in complex situations, such as the stock market. According to the results of the multi-agent simulation and theoretical proofs, the combined model owns the highest accuracy of stock price prediction and results in more profits than other single models. Accordingly, we can conclude that this model would be a suitable and powerful method in directing investment decisions. This newly presented model functions similarly to the prediction system, which not only exhibits a high accuracy rate but also has an effect on guiding operations in the stock market due to the introduction of intelligent agents.