Stock trading using neural networks

• Main Author: Li, Bofeng
• Other Authors: Wang Lipo
• Format: Final Year Project
• Language: English
• Published: 2017
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/72071
• Institution: Nanyang Technological University

This project is aiming to use artificial neural network to predict and analyze the trend of stock price. Stock is a kind of securities, in order to raise the company funds, stock company will issue stock share to the investors and the stock share works as part of the company's capital ownership certificate, and the investors who become shareholders are able to obtain dividends, and also they can share the company's growth or the profits from the trading market fluctuations. But on the other hand, the shareholders also need to share the risks associated with the operation of the company. For this project China stocks market has been used as testing target, as market is big and I am more familiar with it. Neural network is a kind of data processing model which is established under the revelation of biological neural network. The neural network is calculated by connecting a large number of artificial neurons to each other, according to the information from outside world to change their own structure, and mainly by adjusting the weight between the neurons to make a model by the input data entered, and then finally it will have the ability to solve some practical problems, in this project, the neural network is used to solve the stock price trend prediction problem. Matlab is a powerful mathematic software, it combines varies functions and features in one developing platform, such functions as matrix calculation, data analysis, dynamic system modeling and simulation, etc. One of a great and powerful product from Matlab is the ‘Simulink’, it is like a virtual simulation tools to modeling a system, widely used in data signal processing, communication system and digital control system. Other than that, Matlab also include other powerful pre-defined functions and tool boxes. And with the help of the neural network tool box from Matlab, I used to two different neural networks on Matlab platform to analyze and predict the stock price trending which are BP network and Elman network.