How are Genetic Algorithms Used in the NeuroShell Trader Professional? (Professional Only)

In the Prediction Wizard, genetic algorithms are used to find:

  1. The best inputs to a neural network
  2. The best parameters for indicators used as inputs to a neural network
  3. The best buy/sell thresholds for neural nets

In the Trading Strategy Wizard, genetic algorithms are used to find:

  1. The best trading rules to use
  2. The best parameters for indicators used in trading rules
  3. The best parameters for indicators used in limit orders, stop orders, and stop limit orders

In order for a genetic algorithm to optimize, it has to have an idea of what fit individuals in the population look like (the fitness function). The fitness function is the Training Objective in the Prediction Wizard, and the Objective in the Trading Strategy Wizard. You will find that these are objectives like “minimize error”, “maximize correlation”, “minimize drawdown”, “maximize net profit”, etc. To a very great extent, your choice of the objective (fitness) function can have a great bearing on how well the optimization goes, and how well the optimized model works later on out of sample data. Unfortunately, there are no universal guidelines.

 

Topics of Interest:

What are Genetic Algorithms?
How Does a Genetic Algorithm Work?
Why is a Genetic Algorithm Better Than Other Optimizers?
What Types of Problems Do Genetic Algorithms Solve?
 

Was this article helpful?

Related Articles