Earthquake Electromagnetic Precursor Anomaly Detection Method Based on EEMD-LSTM-IF

Numerous studies have shown that earthquakes may be accompanied by anomalous disturbances in the geomagnetic field, so accurate and rapid monitoring of geomagnetic anomalies is necessary. An anomaly detection method based on the combination of Ensemble Empirical Mode Decomposition (EEMD), LSTM neural network and Isolation Forest (IF) algorithm is proposed to detect anomalies of electromagnetic time-series data before an earthquake. First, the decomposition and reconstruction of geomagnetic data using the EEMD algorithm can effectively remove the noise and trend terms in the geomagnetic field data. Then, after data prediction by the LSTM neural network model, outlier determination of the predicted values was performed using the isolated forest algorithm. The results show that the method can effectively detect the electromagnetic anomalies before the earthquake.