AI-Feynman Documentation

Install AI-Feynman
master
AI-Feynman is supported only on Linux and Mac environments. It is recommended to use a fresh virtual environment.
1. Create and activate a virtual environment:
  virtualenv -p python3 feyn source feyn/bin/activate
2. Install dependencies:
  pip install numpy
3. Install the package:
  pip install aifeynman
4. Important: Before running the code, you must compile the Fortran files used for the brute force module by running:
  ./compile.sh
```
virtualenv -p python3 feyn
source feyn/bin/activate
pip install numpy
pip install aifeynman
./compile.sh
```
https://github.com/sj001/ai-feynman
Run AI-Feynman via Python API
master
You can use the aifeynman.run_aifeynman function to perform symbolic regression on a data file.
Parameters:
- pathdir: Path to the directory containing the data file.
- filename: The name of the data file.
- BF_try_time: Time limit for each brute force call (default: 60).
- BF_ops_file_type: File containing the symbols for brute force (default: "14ops.txt").
- polyfit_deg: Maximum degree of the polynomial tried (default: 4).
- NN_epochs: Number of epochs for neural network training (default: 4000).
- vars_name: A list of strings representing the names of the variables in the data file (including the output variable), in the same order as they appear in the file.
- test_percentage: Percentage of input data to set aside for testing (default: 0).
Data Format: The input file must be a text file where each column contains numerical values for each variable (dependent and independent). Columns can be separated by spaces, commas, or tabs.
Output: Results are saved in a directory named results as solution_{filename}. Each row in the solution file represents a point on the Pareto frontier and contains:
- Mean logarithm (base 2) of the error (average error in bits).
- Cumulative logarithm (base 2) of the error.
- Complexity of the equation (in bits).
- Error of the equation on input data.
- The symbolic expression of the equation.
Note: If test_percentage > 0, the error on the test set is added as the first value in each row.
```
import aifeynman

# Download example data
aifeynman.get_demos("example_data")

# Run symbolic regression
aifeynman.run_aifeynman(
    "./example_data/", 
    "example1.txt", 
    60, 
    "14ops.txt", 
    polyfit_deg=3, 
    NN_epochs=500
)
```
https://github.com/sj001/ai-feynman
Run AI-Feynman via Command Line
master
You can call AI-Feynman from the terminal using the ai_feynman_terminal_example.py script. Use the --help flag to see all available parameters.
Example usage:
```
python ai_feynman_terminal_example.py --pathdir=../example_data/ --filename=example1.txt
```
```
python ai_feynman_terminal_example.py --pathdir=../example_data/ --filename=example1.txt
```
https://github.com/sj001/ai-feynman

What's inside AI-Feynman