Search

Browse

• Communities
• Authors
• Year
• Titles
• Subjects

My Space

• Login

Forecasting the stock market trend using a neuro-fuzzy approach (ANFIS)

Vlachos Stavros

Full record

---

Summary

Predicting stock prices or/and indices trend is certainly one of the most important issues in the financial sector and has become one of the major concerns of investors and shareholders, as accurate and authentic forecasts of stock prices or/and indices trend have attractive benefits and profitability advantages while inaccurate and unreliable forecasts can have irreparable consequences. Therefore, it is important to provide an accurate and efficient model for their prediction. The given difficulty is that market volatility, which is non-linear, must be included in forecasting models while factors such as complexity and noisy market environment must be incorporated into them.The purpose of this thesis is to develop a comprehensive Adaptive Neuro-Fuzzy Inference System (ANFIS) for predicting the next day's stock market general index trend as accurately as possible from historical data on stock closing prices. Our approach, based on ANFIS, is justified by the uncertainty and complexity of stock markets that require the mixing of human expertise and mathematical models, adapting to changes and integrating various factors into stock price forecasts. The adaptive neuro-fuzzy system offers a viable alternative to recording stock market behavior: neural networks are used to identify patterns and adapt to meet changing environments and fuzzy systems are used to integrate human knowledge and to perform inference and decision making. The integration of these two with some derivative-free optimization techniques leads to neuro-fuzzy and soft computing based approaches. In the present study the stock market used is the Athens Stock Exchange (ASE).

Available Files

Services