Solar Flare Predictive Features Derived from Polarity Inversion Line Masks in Active Regions Using an Unsupervised Machine Learning Algorithm

The properties of the polarity inversion line (PIL) in solar active regions (ARs) are strongly correlated to flare occurrences. The PIL mask, enclosing the PIL areas, has shown significant potential for improving machine-learning-based flare prediction models. In this study, an unsupervised machine-...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:The Astrophysical journal Ročník 892; číslo 2; s. 140 - 148
Hlavní autoři: Wang, Jingjing, Zhang, Yuhang, Hess Webber, Shea A., Liu, Siqing, Meng, Xuejie, Wang, Tieyan
Médium: Journal Article
Jazyk:angličtina
Vydáno: Philadelphia The American Astronomical Society 01.04.2020
IOP Publishing
Témata:
Algorithms
Astrophysics
Classification
Forest management
Machine learning
Masks
Polarity
Prediction models
Principal components analysis
Solar active regions
Solar activity
Solar activity regions
Solar flares
Space weather
Unsupervised learning
ISSN:0004-637X, 1538-4357
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:The properties of the polarity inversion line (PIL) in solar active regions (ARs) are strongly correlated to flare occurrences. The PIL mask, enclosing the PIL areas, has shown significant potential for improving machine-learning-based flare prediction models. In this study, an unsupervised machine-learning algorithm, Kernel Principle Component Analysis (KPCA), is adopted to directly derive features from the PIL mask and difference PIL mask, and use those features to classify ARs into two categories-non-strong flaring ARs and strong-flaring (M-class and above flares) ARs-for time-in-advance from one hour to 72 hr at a 1 hr cadence. The two best features are selected from the KPCA results to develop random-forest classifiers for predicting flares, and the models are then evaluated and compared to similar models based on the R value and difference R value. The results show that the features derived from the PIL masks by KPCA are effective in predicting flare occurrence, with overall better Fisher ranking scores and similar predictive statistics as the R value characteristics.
Bibliografie:AAS19423
The Sun and the Heliosphere
ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0004-637X
1538-4357
DOI:10.3847/1538-4357/ab7b6c