Earthquake Forecast as a Machine Learning Problem for Imbalanced Datasets: Example of Georgia, Caucasus

Abstract

In this article, we considered the problem of M≥3 earthquake (EQ) forecasting (hindcasting) using a machine learning (ML) approach, using experimental (training) time series on monitoring water-level variations in deep wells as well as geomagnetic and tidal time series in Georgia (Caucasus). For such magnitudes’, the number of “seismic” to “aseismic” days in Georgia is approximately 1:5 and the dataset is close to the balanced one. However, the problem of forecast is practically important for stronger events—say, events of M≥3.5—which means that the learning dataset of Georgia became more imbalanced: the ratio of seismic to aseismic days for in Georgia reaches the values of the order of 1:20 and more. In this case, some accepted ML classification measures, such as accuracy leads to wrong predictions due to a large number of true negative cases. As a result, the minority class, here—seismically active periods—is ignored at all. We applied specific measures to avoid the imbalance effect and exclude the overfitting possibility. After regularization (balancing) of the training data, we build the confusion matrix and performed receiver operating classification in order to forecast the next day probability of M≥3.5 earthquake occurrence. We found that the Matthews’ correlation coefficient (MCC) is the measure, which gives good results even if the negative and positive classes are of very different sizes. Application of MCC to observed geophysical data gives a good forecast of the next day M≥3.5 seismic event probability of the order of 0.8. After randomization of EQ dates in the training dataset, the Matthews’ coefficient efficiency decreases to 0.17.

Earthquake forecast by imbalance machine learning using geophysical predictors

“…In order to correctly assess the effectiveness of ML algorithms in the case of strong imbalance between majority and minority classes it is recommended to use Matthews correlation coefficient MCC or F1 score measure [Matthews, 1975;Chicco and Jurman, 2020;Chicco et al, 2021;Chelidze et al, 2022]:…”

“…As was expected, the Accuracy assessment seems to be too optimistic. This is the result of strong imbalance in the input data -namely, to large values of TN [Chelidze et al, 2022]. As a result, the minority class (strong EQs) is practically ignored in Accuracy assessment.…”

Geophysical Reactions to Remote 2022 Tonga Eruption and to Türkiye Earthquakes in Georgia (Caucasus): Hydrogeology, Geomagnetics and Seismicity

“…WL monitoring network in Georgia includes several deep wells, drilled in confined sub-artesian aquifer (Figure 1), here we use the data of the stations with the most systematic records. The water data are presented either in the pressure units KPa (Figure 2 Magnetometer Model LGI and it accomplishes non-stop registration of X, Y, Z elements variations [10]. The absolute values of components for January 2022 are X-23822.4 nT; Y-2768.3 nT; Z-43543.3 nT.…”

Seismic damage prediction of RC buildings using machine learning

“…The number of datasets of each class should be evenly distributed so that the predictive ML models is not biased toward the majority class and ignored the minority class. 46 In this study, the training datasets are prepared using the simulation results. The advantage of using simulation results is the utilization of amplitude scaling methods of the earthquake wave to increase the datasets for moderate to collapse damage grade, which is explained in the subsequent section.…”

“…Moderate and strong earthquakes are rare events and hence the real‐world datasets are imbalanced and have majority and minority classes. The number of datasets of each class should be evenly distributed so that the predictive ML models is not biased toward the majority class and ignored the minority class 46 . In this study, the training datasets are prepared using the simulation results.…”