My Latest Test Results Of Predicting Earthquake Using Machine Learning CSV Data Feb1990 to Jan 2024

[Jim Smith]

My Latest Test And Development Results Of Predicting Earthquakes Using Machine Learning With CSV Data Feb1990 to Jan 2024 To Try to Predict Feb 2024 and Mar 2024.

The type of machine learning I use here is called a Random_Tree_Regressor. The AI tries to predict something from a lot of unknowns involved. I used about 300 Trees (similar to 300 generations of learning) with a VERY limited amount of CSV Data. 

I find this method of Machine Learning interesting in that, this Machine Learning Method is trying to predict something with a lot of unknowns by using generations of learning trees made in parallel, and it tries to learn which one is most accurate from Limited CSV Data in the M6 to M9 significant range of earthquakes. 

As I mentioned before, use this at your risk. It is subject to change for higher accuracy. After multiple tries to align the test results, you can see the results on a map. The map show numbers 1 to 60 for each day sequence it tried to predict for Feb and Mar 2024.

To get higher accuracy, you would probably need ALL Earthquake Data available for the last 100 years with about 1000000 learning trees, using a Supercomputer or Quantum Computer to better predict the learning trees developed from that Data. What I show here is the very basic hypothetical, and very basic approach using AI, to try to predict something hard to predict with a lot of unknowns.

Note: I worked with the PyCharm AI to obtain the Python Program I used here. Without the AI, it might have taken me about 20+ years to get this far in Machine Learning, especially when trying to predict something with a lot of unknowns. But this only took me a few days to develop...