(This is an archive of the first version of the model utilized in 2018. It has subsequently been superseded by version two starting in 2019.)

Source: United States Geological Survey (USGS)

Training Set: 1973-01-01 to 2016-12-31

Test Set: 2017-01-01 to 2017-12-31

Update: 2018-01-01 to 2018-02-28

Seed Set: Last 1 Day (Previous 24 Hours)

Temporal Resolution: 1 Day (24 Hours)

Spatial Resolution: 2 Degrees Latitude and 2 Degrees Longitude

Details: Earthquakes were reorganized into tensors, with each slice representing a single day and consisting of a matrix with the rows and columns representing approximate latitudes and longitudes. Each position in the matrix was filled with a magnitude determined by converting all the earthquakes occurring at that approximate location and time from logarithmic scale magnitudes into linear energies, summing them together, and then converting the resulting value back into the logarithmic scale.

Long Short Term Memory (LSTM) Recurrent Neural Network (RNN)

Hidden Layers: 5

Neurons Per Hidden Layer: 512 nodes

Timesteps: 1

Epochs of Training: 1

Loss Function: Asymmetric (Exponential Or Logarithmic)

Library: Keras with Theano backend

Notes: Model is fully Stateful and utilizes an Online Learning schedule. Utilizes Layer Normalization and Residual Skip Connections as well as Golden Ratio scaling of the activation functions, and Golden Ratio Conjugate scaling of the gradient norms. No dropout or stochastic timeskip.

- 0.0-0.5: 3075100
- 0.5-1.0: 22426
- 1.0-1.5: 15027
- 1.5-2.0: 14846
- 2.0-2.5: 13726
- 2.5-3.0: 8186
- 3.0-3.5: 7389
- 3.5-4.0: 5913
- 4.0-4.5: 7375
- 4.5-5.0: 367
- 5.0-5.5: 1
- 5.5-6.0: 0
- 6.0-6.5: 0
- 6.5-7.0: 0
- 7.0-7.5: 0
- 7.5-8.0: 0
- 8.0-8.5: 0
- 8.5-9.0: 0
- 9.0-9.5: 0

- A True Positive is scored when the magnitude prediction is within +/-1.0 of the actual earthquake event magnitude, and the actual earthquake event magnitude is greater than 0.0.
- A True Negative is scored when the magnitude prediction is within +/-1.0 of the actual earthquake event magnitude, and the actual earthquake event magnitude is exactly 0.0.
- A False Positive is scored when the magnitude prediction is outside of +/-1.0 of the actual earthquake event magnitude, and the actual earthquake event magnitude is less than the predicted magnitude.
- A False Negative is scored when the magnitude prediction is outside of +/-1.0 of the actual earthquake event magnitude, and the actual earthquake event magnitude is greater than the predicted magnitude.

- True Positive (TP) = 21134.0
- True Negative (TN) = 5912431.0
- False Positive (FP) = 49833.0
- False Negative (FN) = 12046.0
- Total Population (TP + TN + FP + FN) = 5995444.0
- Positive (TP + FN) = 33180.0
- Negative (FP + TN) = 5962264.0
- Predict Positive (TP) = 70967.0
- Predict Negative (TP) = 5924477.0
- Prevalence (PRE = positive / total_pop) = 0.00553420230428
- True Positive Rate OR Sensitivity OR Recall (TPR = TP / positive) = 0.636949969861
- False Negative Rate OR Miss Rate (FNR = FN / positive) = 0.363050030139
- False Positive Rate OR Fall-Out (FPR = FP / negative) = 0.00835806666729
- True Negative Rate OR Specificity (TNR = TN / negative) = 0.991641933333
- Accuracy (ACC = (TP + TN) / total_pop) = 0.989678996251
- Positive Predictive Value OR Precision (PPV = TP / predict_positive) = 0.297800386095
- False Discovery Rate (FDR = FP / predict_positive) = 0.702199613905
- False Omission Rate (FOR = FN / predict_negative) = 0.00203325964469
- Negative Predictive Value (NPV = TN / predict_negative) = 0.997966740355
- Positive Likelihood Ratio (PLR = TPR / FPR) = 76.2078115928
- Negative Likelihood Ratio (NLR = FNR / TNR) = 0.366110001942
- Diagnostic Odds Ratio (DOR = PLR / NLR) = 208.155502959
- Predictive Accuracy (PV = TP / (TP + FP + FN)) = 0.25458663101
- F1-Score = 0.405849424371
- Matthews Correlation Coefficient = 0.431180728834
- Informedness = 0.628591903194
- Markedness = 0.29576712645

- True Positive (TP) = 95.0
- True Negative (TN) = 0.0
- False Positive (FP) = 0.0
- False Negative (FN) = 1279.0
- Total Population (TP + TN + FP + FN) = 1374.0
- Positive (TP + FN) = 1374.0
- Negative (FP + TN) = 0.0
- Predict Positive (TP) = 95.0
- Predict Negative (TP) = 1279.0
- Prevalence (PRE = positive / total_pop) = 1.0
- True Positive Rate OR Sensitivity OR Recall (TPR = TP / positive) = 0.0691411935953
- False Negative Rate OR Miss Rate (FNR = FN / positive) = 0.930858806405
- False Positive Rate OR Fall-Out (FPR = FP / negative) = 0
- True Negative Rate OR Specificity (TNR = TN / negative) = 0
- Accuracy (ACC = (TP + TN) / total_pop) = 0.0691411935953
- Positive Predictive Value OR Precision (PPV = TP / predict_positive) = 1.0
- False Discovery Rate (FDR = FP / predict_positive) = 0.0
- False Omission Rate (FOR = FN / predict_negative) = 1.0
- Negative Predictive Value (NPV = TN / predict_negative) = 0.0
- Positive Likelihood Ratio (PLR = TPR / FPR) = 0
- Negative Likelihood Ratio (NLR = FNR / TNR) = 0
- Diagnostic Odds Ratio (DOR = PLR / NLR) = 0
- Predictive Accuracy (PV = TP / (TP + FP + FN)) = 0.0691411935953
- F1-Score = 0.129339686862
- Matthews Correlation Coefficient = 0
- Informedness = -0.930858806405
- Markedness = 0.0