At AiStockcharts.com a neural network is used as a non-linear statistical data modeling tool that also incorporates an adaptive learning response to the ever-changing stock market conditions. Neural networks are used to model complex relationships between inputs and outputs in order to find patterns in data. You might ask why do we need a neural network instead of some other mathematical tool. The answer is that the performance of a group of stocks in the stock market does not follow any known statistical distribution such as a normal (bell) curve. In fact, most statistical analysis tools cannot be used since distribution transformation does not work in most cases. We are forced to use non-parametric statistical analysis to analyze technical indicators and their relationship to stock performance.

Click here for additional information on neural networks and artificial intelligence, including diagrams of neural networks.

AiStockCharts.com continually analyzes the performance of all S&P 500 stocks based on a growing number of technical indicators. The performance is input into an artificial neural network. The network is adaptive and learns which technical indicators have worked best for particular stocks then leverages that information to make decisions. One decision could be to create a new technical indicator based on what has worked recently in the stock market. Another decision might be to increase the update frequency for certain technical indicators based on volatility or a sudden change in the performance of similar indicators. Decisions are automatically analyzed so that the network learns from its mistakes.

AiStockcharts.com groups similar technical indicators based in part on performance. Indicators that have been found to be historically insignificant are ignored (however, not forever).