Machine learning helps forecast time to earthquake


Researchers at Los Alamos National Laboratory recently applied machine-learning to forecast earthquakes along Cascadia, which is located within the western region of North America.

After analyzing Cascadia data, the team of researchers discovered the megathrust broadcasts a constant tremor, a fingerprint of the fault’s displacement. It is interesting to note here that the team found a link between the loudness of the fault’s acoustic signal and its physical changes, said a report.

“Cascadia’s behavior was buried in the data. Until machine learning revealed precise patterns, we all discarded the continuous signal as noise, but it was full of rich information. We discovered a highly predictable sound pattern that indicates slippage and fault failure,” Paul Johnson, Los Alamos scientist, was quoted as saying.

“We also found a precise link between the fragility of the fault and the signal’s strength, which can help us more accurately predict a megaquake,” Johnson added.

What machine learning actually does is to chomp massive seismic data sets to obtain distinct patterns. This is done by the way of learning from self-adjusting algorithms to generate decision trees that choose and retest a series of questions and answers.

Get latest updates in your inbox, subscibe to our daily newsletter.