Predicting Solar Magnetic Activity from <i>S</i> _ph and Seismic Parameters Using Random Forest Regression

Abstract We investigate the potential of using the photometric magnetic proxy S ph and seismic parameters, such as the frequency of maximum power ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:msub> <mml:mrow> <mml:mi>ν</mml:mi> </mml:mrow> <mml:mrow> <mml:mi mathvariant="normal">max</mml:mi> </mml:mrow> </mml:msub> </mml:math> ) and the large frequency separation (Δ ν ), derived from Solar and Heliospheric Observatory/Variability of Solar Irradiance and Gravity Oscillations observations to predict the 10.7 cm solar radio flux, a widely used index of solar magnetic activity. A random forest regression model is trained and tested on time series divided into multiple temporal subsets and input parameter combinations. The model achieves strong predictive performance ( R 2 &gt; 0.92) across configurations and significantly outperforms a classical linear regression model. Our results show that S ph effectively captures long-term variations, while the seismic amplitude parameter <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:msub> <mml:mrow> <mml:mi>H</mml:mi> </mml:mrow> <mml:mrow> <mml:mi>max</mml:mi> </mml:mrow> </mml:msub> </mml:math> is more responsive to short-term fluctuations. Combining S ph with the full set of seismic parameters yields the highest accuracy and offers a promising approach for diagnosing activity in other solar-like stars where direct magnetic field measurements are infeasible.