Skip to content

In response to Siraj's Week #4 challege : Predict the magnitude of an earthquake, given the USGS dataset

Notifications You must be signed in to change notification settings

nacmonad/nn-earthquakes

Repository files navigation

nn-earthquakes

In response to Siraj's Week #4 challege : Predict the magnitude of an earthquake, given the USGS dataset.

We are given a NN, modify it to predict the magnitudes of earthquakes. The base dataset

Steps:

  1. Clean Data

  2. Normalize Data -- Using minmax scaling z= (x-min(x))/(max(x)-min(x)) (Done in R)

  3. Add extra features and repeat 1 & 2 (incomplete)

Neural Network source: https://github.com/sibyjackgrove/Earthquake_predict

  1. Train

  2. Predict -- Use random_search & grid_search for hyperparameter tuning (http://scikit-learn.org/stable/auto_examples/model_selection/randomized_search.html)

A challenge to myself is to increase the predictive capabalities of the net from the base USGS dataset, by adding the lunar-distance, and solar flare activity. But I'm having troubles...

https://www.fourmilab.ch/earthview/pacalc.html - lunar apogee/perigrees

According to USGS, a link between solar activity and earthquakes has not been established : https://www2.usgs.gov/faq/categories/9830/3282

However a look at the weights in the NN might be able to tell us if a machine-learning algorithm finds it to be relevant data for the prediction of earthquakes magnitues....

About

In response to Siraj's Week #4 challege : Predict the magnitude of an earthquake, given the USGS dataset

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published