Font Size: a A A

Stochastic Neural Network Forecasting Model And Financial Market Volatility And Correlation Research

Posted on:2016-04-23Degree:MasterType:Thesis
Country:ChinaCandidate:H Y MoFull Text:PDF
GTID:2309330470455610Subject:Statistics
Abstract/Summary:PDF Full Text Request
With the development of financial market, forecasting the stock market volatility has become an important financial issue. A key element of financial forecasting is to build a model mining financial data in the internal correlation. Artificial Neural Net-work (ANN) is a non-linear statistical modeling and decision-making methods, widely used in prediction and analysis of financial time series. Based on the long memory and fat tails of stock index return, we construct Stochastic Time Strength Neural Network (STSNN) and Exponent Back Propagation Neural Network (EBPNN). In STSNN mod-el, stochastic time strength function gives a weight for each historical data and makes the model have the effect of random movement. In EBPNN model, information is not only processed locally in each neural unit by computing the dot product between its in-put vector and its weight vector, but also processed by adding the dot product between its exponential type function of the input vector and its corresponding new weight vec-tor.In empirical analysis section, SSE, SZSE, HSI, DJIA, IXIC and S&P500with d-ifferent selected volatility degrees and return scaling cross-correlations are predicted by STSNN model. And the empirical research is performed in testing the forecasting effect of EBPNN model by applying different exponential parameter functions and a compar-ison to back propagation neural network (BPNN). The empirical analysis shows the STSNN and EBPNN model are advantageous in increasing the forecasting precision.
Keywords/Search Tags:neural network forecasting, stochastic time strength function, exponen-tial function, stock index, correlation coefficient
PDF Full Text Request
Related items