Included in this pull request are two new symbolic regression examples, that make it easier to run your own symbolic regression using your own tabular datasets. They are far more better than the older ones I had created.
The first notebook will walk you step by step through a mostly automated process pointing out where and what to configure.
The second notebook includes code to run a Multidemic version of the training, to test Island-Migration evolution models.
Extra features found in this update are mainly in the notebooks, but include:
data dictionary based pre-processing, just provide it a data dictionary, and the data pipeline adjusts itself
improved linker functions enabling the ability for example to average results over many genes
improved performance, through employing a jupyter notebook friendly version of multiprocessing
options to test normalizing your input data, to see if this improves the regressions
options to test including K-means clustering to your input data
more example datasets, and the needed dictionaries, to help you practice
Defaults included are fairly good, and it should do well with minimal adjustment, apart from giving it new data.
ShuhuaGao
commented
1 year ago
Included in this pull request are two new symbolic regression examples, that make it easier to run your own symbolic regression using your own tabular datasets. They are far more better than the older ones I had created.
v1.0.0_YourSymbolicRegression-Master.ipynb v1.0.3_YourSymbolicRegression-Multidemic.ipynb
The first notebook will walk you step by step through a mostly automated process pointing out where and what to configure. The second notebook includes code to run a Multidemic version of the training, to test Island-Migration evolution models.
Extra features found in this update are mainly in the notebooks, but include:
Defaults included are fairly good, and it should do well with minimal adjustment, apart from giving it new data.