A Novel Fuzzy Time Series Model for Stock Market Index Analysis using Neural Network with Tracking Signal Approach

Abstract

Objectives: To present a novel Fuzzy Time Series Neural Network (FTSNN) with Tracking Signal (TS) approach for forecasting the closing index of the stock market. Methods/Statistical Analysis: A novel approach strives to adjust the number of hidden neurons of a Multi-Layer Feed Forward Neural Network (MLFFNN) model. It uses the Tracking Signal (TS) and rejects all models which result in values outside the interval of [-4, 4]. Findings: The effectiveness of the proposed approach is verified with one step ahead of Bombay Stock Exchange (BSE100) closing stock index of Indian stock market and Taiwan Stock Exchange Stock Index (TAIEX). This novel approach reduces the over-fitting problem, reduces the neural network structure and improves forecasting accuracy. In addition, the presented approach has been tested on standard NN3 (Neural Network 3) forecasting competition time series dataset and this approach out performs the various models tested with the NN3 forecasting competition. Applications/Improvements: The proposed approach can be applied to different types of neural network for forecasting closing stock index/price of stock market data.