Genetic Programming Fitness Choices for Quantitative Signals

Summary

The document identifies the fitness-function setting used by a genetic programming module and lists its available objectives. These include information coefficient to information coefficient ratio, mutual information, long-only Sharpe, volatility and return measures, and long-short Sharpe, volatility and return measures. The default is the information-coefficient-to-information-coefficient-ratio option.

This list shows that candidate expressions can be evaluated using different statistical or portfolio-oriented criteria, depending on whether the aim is long-only or long-short performance. However, the document does not define how any metric is calculated, what the returned values mean, or how the module handles scaling, costs, or constraints. It is therefore a menu of fitness-function names rather than a guide to interpreting their values or comparing genetic programming results.

Key ideas

• The module requires a fitness-function choice and defaults to an ICIR objective.
• Available objectives include mutual information and return, volatility, or Sharpe measures.
• Several objectives distinguish long-only evaluation from long-short evaluation.
• The document names the choices but does not explain their formulas or output interpretation.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.