seisclass 0.1.7

Seismic event classification package

seisclass

Seismic event classification package for identifying natural and non-natural earthquakes.

Overview

seisclass is a Python package designed to classify seismic events using deep learning models. It provides one main function:

• check_seed: Analyzes a seismic data file (SEED format) with corresponding phase file, returning a comma-separated result string: event_type,earthquake_prob,explode_prob,collapse_prob

For detailed information about the program and to cite it in your research publications, please refer to the following papers:

[1] Jia, L., Chen, H., & Xing, K. (2022). Rapid classification of local seismic events using machine learning. Journal of Seismology, 26(5), 897-912.

[2] Jia, L., Chen, S., Li, Y., & Zheng, P. (2025). A Semisupervised Seismic Events Classifier Based on Generative Adversarial Network. Seismological Research Letters, 96(3), 2039-2051.

Git-Repository: [https://github.com/epnet2018/]

Usage

Basic Usage

from seisclass import check_seed
# Analyze a SEED file with corresponding phase file
result = check_seed('path/to/seed/file', 'path/to/phase/file')
print(result)

phase file format

The phase file should be in the following format:

station,channel,time,phase

where station is the station code, channel is the channel code, time is the arrival time of the phase, and phase is the phase type (e.g., P, S).

Net_code Sta_code Loc_id Chn_code Phase_name Phase_time Phase_time_frac Resi Mag_val Distance Azi

XX XXXXX 00 HHZ P 2025-09-20 03:20:39 7600 -1.903390 2.391120 53.238400 339.952000

YY YYYYY 00 HHZ P 2025-09-20 03:20:46 4100 -2.548810 2.651560 104.083000 90.712900

Installation

You can install the package from PyPI:

pip install seisclass

Or install from source:

pip install .

Dependencies

• numpy
• obspy
• tensorflow
• keras
• joblib
• scikit-learn
• pandas

Advanced Usage

# Specify a different model
result = check_seed('path/to/seed/file', 'path/to/phase/file', model_str='251111nw')

Testing

To run the tests:

pytest

License

This project is licensed under the MIT License - see the LICENSE file for details.

Authors

• Jia Luozhao - [lezhao.jia At gmail.com](mailto:lezhao.jia At gmail.com)