Speaker
Description
The planetary Kp index is one of the main indicators used to monitor geomagnetic activity. The official Kp index is reported only at 3 hour intervals, making it difficult to resolve rapid changes in geomagnetic activity driven by solar wind variability. In this work, we present a machine learning model developed to estimate the Kp index with a temporal resolution of one minute using interplanetary solar wind data. The model was trained using historical Kp values together with solar wind parameters, including the interplanetary magnetic field, solar wind speed, and proton density. A Random Forest regression algorithm was selected because it efficiently captures the nonlinear relationship between solar wind conditions and geomagnetic activity. The proposed model provides a continuous estimate of Kp, allowing a more detailed visualization of geomagnetic variations than the official 3 hour index. The system can also operate with near-real-time solar wind measurements, making it suitable for space weather monitoring and future forecasting applications. These results demonstrate the potential of machine learning techniques to improve the temporal representation of geomagnetic activity and to support operational space weather studies.