Stanford University researchers have developed Fingerprint And Similarity Thresholding (FAST), an algorithm that could transform how seismologists detect temblors that are not strong enough to register as earthquakes when analyzed by conventional methods.
Monitoring microquakes could help scientists predict how frequently, and where, larger quakes are likely to occur.
The FAST technique takes all recorded data from a seismic station and divides the continuous signal into segments of a few seconds each, and then compresses the signals into compact representations, or "fingerprints," for rapid processing. The fingerprints are then sorted into separate groups based on their similarities.
"Tests we have done on a six-month data-set show that FAST finds matches about 3,000 times faster than conventional techniques," says Stanford professor Greg Beroza. "Larger data-sets should show an even greater advantage."
The researchers created FAST using techniques from data mining and machine learning. FAST's scalability stems from the use of locality-sensitive hashing (LSH), which is "a widely used technique for identifying similar items in large data-sets," says Stanford researcher Karianne Bergen.
The researchers are scaling up the FAST algorithm to analyze data collected across longer periods of time, from multiple seismic stations, and in more challenging scenarios.
From Stanford Report
View Full Article
Abstracts Copyright © 2015 Information Inc., Bethesda, Maryland, USA