Skip to content

Classify ground motion waves into earthquakes or blasts using traditional Machine Learning algorithms.

License

Notifications You must be signed in to change notification settings

VedangW/GroundMotionClassifier

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

20 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

GroundMotionClassifier

A project to differentiate between earthquakes and blasting waves using Support Vector Machines.

Prerequisites:

To run this project, you would need a linux-based operating system (Ubuntu or Fedora would work best).

The code is written in Python 2.7.12+, but any version of Python 2 would work.

You would also need the following installed in your system:

  • Scipy
  • Numpy
  • Matplotlib
  • Scikit-Learn
  • Peakutils
  • Plotly

These can be downloaded using a download manager such as pip.

Install pip:

sudo apt-get install python-pip

Install any of the dependencies with pip. For eg,:

pip install scikit-learn
pip install numpy

Running the code:

The feature vector is stored in store.txt present in isrsvm/PS/Code. To create a new feature vector (while erasing the previous one):

sh run.sh

To test the working of any module, you can simply compile it with Python 2 with the appropriate command-line arguments. Check in the comments in the relevant file to know the command-line arguments. For eg.:

python Seismogram.py Kachchh pitsa001.044
python rsp.py /path/to/PS/Datasets/Surendranagar pitsa001.003 r

To train the classifier and plot the decision boundary along with the scatterplot, compile the classifier file. This however should be done after creating the feature vector:

python classifier.py

Datasets:

The datasets are present in isrsvm/PS/Datasets.

Note: These datasets are owned by the Institute of Seismological Research, Gandhinagar, India.

About

Classify ground motion waves into earthquakes or blasts using traditional Machine Learning algorithms.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published