A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf
.
Identification of synoptic weather types over Taiwan area with multiple classifiers
2018
Atmospheric Science Letters
In this study, a novel machine learning approach was used to classify three types of synoptic weather events in Taiwan area from 2001 to 2010. We used reanalysis data with three machine learning algorithms to recognize weather systems and evaluated their performance. Overall, the classifiers successfully identified 52-83% of weather events (hit rate), which is higher than the performance of traditional objective methods. The results showed that the machine learning approach gave low false alarm
doi:10.1002/asl.861
fatcat:ijzm363febhk5icd4oynfftr5u