Font Size: a A A

Stock Price Forecasting Based On Artificial Neural Network Model

Posted on:2005-07-29Degree:MasterType:Thesis
Country:ChinaCandidate:X W WuFull Text:PDF
GTID:2206360122997197Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
Stock market is a highly complicate nonlinear system. Its variation has its own regulation, but also is influenced by many other factors such as market, economy and non-economy. While traditional prediction techniques are based on statistics and meet difficulties in stock market analysis, neural networks enjoy the virtue of self-organization and adaptability and can learn the economical knowledge form historical data. So it is suited to solve problems in stock market prediction.A stock price prediction model is founded on Artificial Neural Network with a hybrid training algorithm in the paper. The model is based on the predictability of stock market and quantitating part of the influence factors in stock market and using the traditional stock technical analysis method.Firstly, it is stated that the stock market is predictable and the in the paper. Influence factors in stock market are listed in the paper. The stock analysis methods and stock market forecast methods home and aboard are expounded.Secondly, the hybrid training algorithm is proposed on the base of the Error Back-propagation algorithm's disadvantage analysis in the paper. The proposed algorithm improve the network convergence speed from the aspects that are the neuron numbers of the hidden layer, the ranges of initial weights and initial threshold, error function and the self-adjusted study rate. The genetic algorithm is imported to make weight evolution computation in the training algorithm in order to avoid plunging the local minimum.Next, the stock price prediction model is proposed on the base of nonlinear time series prediction theory. The model is based on the three-layer feed-forward network with a hybrid algorithm. Input variables in the model are selected according to the relativity of the stock price influence factors. According to the model, the stock price prediction system is designed and developed.Finally, the validity of the model is tested by the stock price prediction experiments whose prediction objective is stock close index or its amplitude with Shanghai A Indexes and ShenFaZhanA.
Keywords/Search Tags:Artificial Neural Network, stock price prediction model, back propagation algorithm, weight evolution, hybrid algorithm
PDF Full Text Request
Related items