Modeling Chaotic Behavior of Stock Indices Using Intelligent Paradigms

Title: Modeling Chaotic Behavior of Stock Indices Using Intelligent Paradigms

Research Question: Can intelligent systems accurately model the chaotic behavior of stock indices?

Methodology: The researchers used four different connectionist paradigms to model the Nasdaq-100 index of Nasdaq Stock Market and the S&P CNX NIFTY stock index. These paradigms included an artificial neural network trained using the Levenberg-Marquardt algorithm, Support Vector Machine (SVM), Difference Boosting Neural Network (DBNN), and a Takagi-Sugeno neuro-fuzzy model. They analyzed 7 years' Nasdaq 100 main index values and 4 years' NIFTY index values to determine the accuracy of each model.

Results: The study found that all four connectionist paradigms could represent the stock indices behavior very accurately. The artificial neural network, SVM, DBNN, and neuro-fuzzy model all provided reliable and efficient techniques for modeling the stock market indices.

Implications: The results suggest that intelligent systems can effectively model the chaotic behavior of stock indices. This has practical implications for investors and traders, as these models could potentially be used to predict stock market trends and make more informed investment decisions. Additionally, the study contributes to the broader field of financial forecasting by demonstrating the effectiveness of different connectionist paradigms in modeling complex, non-linear systems.

Link to Article: https://arxiv.org/abs/0405018v1 Authors: arXiv ID: 0405018v1