Neural Network modeling

Neural Networks (NN) has during the last few years got a lot of attention in all kind of industries. Traders have started to use NN’s for a wide variety of purposes – everything from predicting volatility, confirming signals or even trying to predict market returns. They can have a significant influence to your trading decision, but the nature of NN’s is very complex and has a steep learning curve.

A neural network is a system of programs and data structures that approximates the operation of the human brain:

A neural network is an interconnected assembly of simple processing elements, units or nodes, whose functionality is loosely based on the animal neuron. The processing ability of the network is stored in the interunit connection strengths, or weights, obtained by a process of adaptation to, or learning from, a set of training patterns. (Kevin Gurney)

In other words, a NN can be thought of as a very complex regression model, where each input is assigned a certain weight. These inputs are connected in parallel in layers, with additional functions based on the problem it is intended to solve. This means that a NN can find relationships in the data that is not apparent to the human eye, thus make better estimations of values than the trader or a mathematical formula can.

Now, how does this apply to your trading? Many clients prefer to confirm their signals with a NN model. Every time your strategy signals to open an order, the order is confirmed with the NN whether to be approved or ignored (or if we measure strength of the signal the position size can be adjusted accordingly). This is done by finding variables that correlate to the decision you (or your strategy) is making, setting up a model to analyze the interrelationships, train the model with data and finally deploy it. By using a NN in your trading, it can help you avoid trades that have a higher probability of going against you, and give you more confidence in the trades that does turn out profitable.

A NN study starts at €400. Please contact us for further discussion.

Contact us