AI Stock Trend - Forecasting stock prices and trends with evolutionary computation
The Stock Market plays a pivotal role in the financial market environment. Moreover a source of financing for corporations, it also has become a means of personal financing for individuals as well. This type of growth foments competition between investors, which is essential in forecasting the market's behavior.
For many years, numerous improvements have been made in the use of mathematical, statistical, and computational techniques in financial market trend forecasting from historical data. The artificial neural network excels as an innovative technique due to its high degree of complex mathematical functions and its stochastic nature.
This solution proposes that researchers allocated in a business environment, develop software that uses MLP (multi-layer perceptron) neural network and a learning algorithm known as backpropagation with the objective of forecasting stock prices and trends.
The key innovation factor of this project is the application of artificial neural networks in the process of forecasting future values of stock prices, and also the use of evolutionary computation in the selection of portfolios (also known as exogenous series) which best correlate with the system's forecast.
Although the forecast is not fail proof, the use of artificial neural networks obtained better results compared to various statistical methods, which generates a certain degree of confidence that the system will assist the investor in his decision making process.
Fruition
The initial result of the project is software called A.I. StockTrend, which can be opened at www.aistocktrend.com . A.I. Stock Trend is still a research project, but is being prepared to support the decision making process of investors in the Brazilian stock market and FOREX.
In both 2009 and 2010, the project got an article published at one of the most reputable computer intelligence conference: The International Joint Conference on Neural Networks (IJCNN) in Atlanta and Barcelona;
The project published another article at the International Conference on Computing in Economics and Finance (CEF2010) in London;
The initial result of the project is software called A.I. StockTrend, which can be opened at www.aistocktrend.com . A.I. Stock Trend is still a research project, but is being prepared to support the decision making process of investors in the Brazilian stock market and FOREX.
In both 2009 and 2010, the project got an article published at one of the most reputable computer intelligence conference: The International Joint Conference on Neural Networks (IJCNN) in Atlanta and Barcelona;
The project published another article at the International Conference on Computing in Economics and Finance (CEF2010) in London;
