The paper presents a methodology, ARGOS (Automatic Regression for Governing Equations), for automatically discovering ordinary differential equations (ODEs) from empirical data. ARGOS integrates denoising techniques, sparse regression, and bootstrap confidence intervals to identify dynamical laws. The method is evaluated on well-known ODEs with varying initial conditions, time series lengths, and signal-to-noise ratios. ARGOS consistently identifies three-dimensional systems with moderately sized time series and high signal quality. Compared to SINDy with AIC, ARGOS achieves higher identification accuracy and efficiency, especially in low to medium-noise conditions. The paper also discusses the limitations and potential improvements, emphasizing the importance of data quality and quantity for accurate model discovery. The results highlight the effectiveness of ARGOS in automating the discovery of mathematical models from data, particularly in complex systems where manual model development is challenging.The paper presents a methodology, ARGOS (Automatic Regression for Governing Equations), for automatically discovering ordinary differential equations (ODEs) from empirical data. ARGOS integrates denoising techniques, sparse regression, and bootstrap confidence intervals to identify dynamical laws. The method is evaluated on well-known ODEs with varying initial conditions, time series lengths, and signal-to-noise ratios. ARGOS consistently identifies three-dimensional systems with moderately sized time series and high signal quality. Compared to SINDy with AIC, ARGOS achieves higher identification accuracy and efficiency, especially in low to medium-noise conditions. The paper also discusses the limitations and potential improvements, emphasizing the importance of data quality and quantity for accurate model discovery. The results highlight the effectiveness of ARGOS in automating the discovery of mathematical models from data, particularly in complex systems where manual model development is challenging.