论文标题
经常性的卷积神经网络有助于预测地震的位置
Recurrent Convolutional Neural Networks help to predict location of Earthquakes
论文作者
论文摘要
我们研究了现代神经网络体系结构对地震中期预测的适用性。我们的基于数据的分类模型旨在预测阈值高于阈值的地震是否在给定$ 10 $ 10 $ 10 $ - $ 60 $ 60美元的给定面积上发生。我们深层的神经网络模型具有反复发作的部分（LSTM），该部分解释了地震与卷积部分之间的时间依赖性，该卷积部分解释了空间依赖性。获得的结果表明，基于神经网络的模型击败基于基线特征的模型，这些模型也解释了不同地震之间的时空依赖性。对于日本地震的历史数据，我们的模型预测，从给定的时刻开始发生地震，$ 10 $ 60美元，幅度$ m_c> 5 $，带有质量指标roc auc $ 0.975 $和pr auc $ 0.0890 $警报。基线方法具有类似的Roc AUC $ 0.992 $，正确的预测数$ 1.19 \ CDOT 10^3 $，并且缺少$ 2.07 \ CDOT 10^3 $地震，但较差的pr auc $ 0.00911 $和虚假警报数量$ 1004 \ cdot 10^3 $。
We examine the applicability of modern neural network architectures to the midterm prediction of earthquakes. Our data-based classification model aims to predict if an earthquake with the magnitude above a threshold takes place at a given area of size $10 \times 10$ kilometers in $10$-$60$ days from a given moment. Our deep neural network model has a recurrent part (LSTM) that accounts for time dependencies between earthquakes and a convolutional part that accounts for spatial dependencies. Obtained results show that neural networks-based models beat baseline feature-based models that also account for spatio-temporal dependencies between different earthquakes. For historical data on Japan earthquakes our model predicts occurrence of an earthquake in $10$ to $60$ days from a given moment with magnitude $M_c > 5$ with quality metrics ROC AUC $0.975$ and PR AUC $0.0890$, making $1.18 \cdot 10^3$ correct predictions, while missing $2.09 \cdot 10^3$ earthquakes and making $192 \cdot 10^3$ false alarms. The baseline approach has similar ROC AUC $0.992$, number of correct predictions $1.19 \cdot 10^3$, and missing $2.07 \cdot 10^3$ earthquakes, but significantly worse PR AUC $0.00911$, and number of false alarms $1004 \cdot 10^3$.
