Table 1 Detection performance.

From: Earthquake transformer—an attentive deep-learning model for simultaneous earthquake detection and phase picking

Model

Pr

Re

F1

Training data

Training size

Ref.

EQTransformer

1.0

1.0

1.0

Global

1.2M

This study

CRED

1.0

0.96

0.98

Global

1.2M

7

DetNet

1.0

0.89

0.94

China

30K

5

Yews

0.84

0.85

0.85

Taiwan

1.4M

4

STA/LTA

0.91

1.0

0.95

—

—

11

  1. Pr, Re, and F1 are precision, recall, and F1-score respectively. EQTransformer and CRED have been trained on STEAD dataset while DetNet and Yews results are based on pre-trained models on different datasets. Recursive STA/LTA algorithm is used here.
  2. Bold values represent the best performance.
Back to article page