Font Size: a A A

An Analytical Study On Stock Price Of Listed Companies Returning To Fundamental Plane Based On Genetic Neural Network

Posted on:2013-09-13Degree:MasterType:Thesis
Country:ChinaCandidate:Z W ZhuFull Text:PDF
GTID:2268330425492611Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
Stock market is no doubt to boost the national economy, especially in the economic context of globalization. Securities markets affect the healthy development of the country’s economic competitiveness directly. In China’s capital market, there is a point of view that the stock performance of listed companies in China has nothing to do with the fundamentals and the price distortion.The stock market is an extremely complex nonlinear system. Neural networks, however, has a strong self-learning, adaptive and nonlinear approximation ability and other characteristics. According to the principle of survival of the fittest genetic algorithm global search, search area is unknown. Genetic algorithms and neural network algorithms can be combined to give full play to their strengths. So that, the paper take genetic algorithms and neural networks to research relation between technology stock prices of listed companies and the fundamentals of the relevance of empirical.This paper analyzes the impact of China’s stock prices of many factors. From the macro-economic trends, performance of listed companies, other factors affect our stock price of these three main factors. The paper establish a neural network model. The model’s input parameters is namely, earnings per share, net return on assets, the flow of capital, GDP and CPI. The model’s output parameter is the stock price.The paper uses BP (Back-Propagation) neural network modeling. According to experiments, the paper confirm the model number of layers and neurons and the weight matrix. The Levenberg-Marquardt algorithm of BP neural network has been improved, reducing the BP neural network training time and improve the accuracy of BP neural network convergence. By genetic algorithms and neural networks, genetic algorithm of BP neural network initial weights and threshold optimization. The paper build a research model to research between shares of listed companies and the fundamentals.Finally, the model use MATLAB to emulate, train and validate. The results showed that the price run away from fundamentals for a long time, but with the market order of norms and the healthy development of the stock market, shares of listed companies tend to gradually return value.
Keywords/Search Tags:Stock price, BP neural network, Genetic algorithm, Fundamentals of listedcompanies
PDF Full Text Request
Related items