Actual Performance from Simple S&P500 Training

In the last table you saw the results from the training errors from 1000 RPROP iterations of each of the S&P 500. Some companies trained better than others. Perhaps the patterns in their price movements were more repetitive. The next step is to take these neural networks, that were trained with historical data, and use them to attempt to predict recent price fluctuations. We will use all 500 companies again and see how they can predict the last 60 days of stock prices. This table summarizes this.

Here you see the percent accuracy for each of the 500 companies over the past 60 days. The percent shows what percent of the days the neural network correctly predicted the direction of the stock. Many cases are less than 50%. However there are some that are above 60%. This is all new data that the neural network never saw before(was not trained with). Generally, the companies that had higher prediction averages were also the ones that trained better.

This is a simple neural network that uses ONLY the historic price of a stock to predict its future direction. This really is just about the most simple prediction neural network you can create. I am going to experiment further with more complex neural networks. But this will be the baseline, to see how much more accurate other network types and input patterns will be. I also hope to add an automated program that will automatically post Encog predictions in real time, for the next day, to see how well it performs on an ongoing basis. Obviously, this is for education purposes ONLY and s should not be used to invest!