Abstract Fast forward interplanetary shocks (FFs) are one of the primary drivers of space weather events. This study presents an automated detection algorithm for FFs based on a multilayer perceptron model, utilizing in situ measurements of interplanetary magnetic fields and solar wind plasma from the Wind spacecraft at 1 AU. The training data set was constructed using FF events from the Harvard‐Smithsonian Center for Astrophysics (CfA) shock list between 1995 and 2019, while the testing data set comprised FF events from a combined catalog of the CfA and Helsinki University shock lists from 2020 to 2024. During the testing period, the method identified 169 FFs, of which 83 matched entries in the existing catalog. Among the remaining 86 uncatalogued events, 67 satisfied the Rankine‐Hugoniot (R‐H) jump conditions, confirming them as novel shocks not previously recorded. Using direct comparison with the composite CfA/Helsinki catalog, the model achieves a recall of 80.58%. Since some apparent false positives in this direct comparison are later supported as physically consistent shocks by the R‐H checks, the combined evaluation against the expanded benchmark yields a recall of 81.62% with false alarm rate = 10.65%. This approach provides a robust tool for the classification and study of interplanetary shock properties, thus enhancing the capability for space weather forecasting.
谭熠 et al. (Wed,) studied this question.