Providing an Intelligent Model Based on an Adaptive Fuzzy Artificial Neural Network for Stock Price Prediction
نویسندگان
1 Department of Accounting, Se.C., Islamic Azad University, Semnan, Iran
2 Department of Accounting, Sar.C., Islamic Azad University, Sari, Iran
3 Department of Accounting, Sar.C., Islamic Azad University, Sari, Iran
doi
چکیده
Investment is one of the most important topics in the economies of all countries, with significant importance for individuals and high-level officials. In this research, we estimate the stock returns of sample companies using an adaptive fuzzy neural network model. In this method, fuzzy logic is used to improve the performance of neural networks by adding the concept of uncertainty. In the present study, initial price, highest and lowest price, closing price, and trading volume variables were injected into the model as input data. By defining and preprocessing data related to listed companies, the data were divided into two categories: training and testing, and in the design of the hybrid network model, 6 input variables and 1 output variable were used. Then, by converting the input data into fuzzy numbers, the basic fuzzy inference system model was designed, and a mathematical model for selecting the optimal stock portfolio was introduced. The results showed that Bank Mellat's stock was placed as the best recommended stock in the trading market. The proposed intelligent method can replace current methods in existing stock price prediction software in the stock exchange to help brokers, and investors can also benefit from the model presented in the present study to improve their decision-making power.
