Stock Investing with Neural Networks

Today I want to talk about a very interesting topic. It’s about Neural Networks used to facilitate stock investing decisions.

Generally, you have a lot of neurons, which build the neural network, in your body. What I am talking about is the artificial neural network. Artificial in so far in that it is made up by bits and bytes. Therefore, it is a software which gets impulses and sends responses.

The basic structure is exactly the same like the one you learned in school. You have an input, then one or several hidden layers and the output.


The hidden layer is the part of the system where the computing is taking place. The computing formulas are put in upfront. But it can only compute what it is given. Thus, the “feed” is very important for getting good results. If good stock investing is the goal then only related and relevant data is needed. Just like in real life, “you can only reap what you sow”.

And just like in real life you earn when you learn. The neural network will only give satisfying results if it gets enough input. After a certain time the neural network will give certain units within the hidden layer more weight and others lesser. The unit-weights are fractions of 100%.

The tester or stock investor must be careful and know when to stop the learning process. Otherwise an overlearning could occur. This means that after having reached the best results with the neural network in terms of delivering adequate stock investing recommendations the output quality sinks. Therefore, the recommendations become worse.

Stock Investing