Feynman Symbolic Regression Formulas for Scientific Discovery

Comprising formulas from the Feynman Symbolic Regression Database, designed to evaluate Symbolic Regression methods for Scientific Discovery. The dataset is part of an 'Easy set' curated by yoshitomo-matsubara and was last updated in March 2024.

Use Cases

• Evaluate Symbolic Regression methods on the Feynman formula set to assess their potential for scientific rediscovery.
• Benchmark the performance of regression algorithms using the curated 'Easy set' of formulas.
• Analyze the properties of formulas and their variables from the Feynman Symbolic Regression Database.

Strengths

• Dataset is part of the SRSD series specifically designed for evaluating Symbolic Regression for Scientific Discovery.
• Formulas are based on the Feynman Symbolic Regression Database, a known resource in physics.
• The dataset page includes a detailed description and an associated arXiv paper (2206.10540).

Limitations

• The number of rows, columns, and sample data are unknown, limiting reproducibility and scale assessment.
• Specific file formats and dataset size are not provided, complicating integration and usage planning.
• The 'Easy set' may represent a simplified subset, potentially limiting challenge for advanced SR methods.

Provenance

Source

Feynman Symbolic Regression Database, extended by yoshitomo-matsubara.

Collection Method

Formulas were reviewed and sampling ranges were designed to be reasonably realistic for evaluation.

Freshness

Last updated on 2024-03-05.