A simple Genetic Program to get in the top 100 for the main Numer.ai competition.
Run the python file after changing the directory to point to the current Numer.ai data
python allornothing.py will produce two files. Genetic Programming discovered the basic models. The output was then used for predictions; one is optimistic where as one is feature neutralized. python symmetric.py will also produce files. One uses a GP Relu function and the other uses a Sigmoid Function to discover a symmetric representation of the data. Predictions are then made using linear algebra.
Initial Upload
Scirpus specializes in Genetic Programming
MIT License