how a scientist established a two-stage solar flare early warning system
The visualization of four features during the existence of an active region. The x-axis represents time and its unit is a sample, where “0” represents the start time of an active region, and the time gap between adjacent times is 1.5 h. The y-axis represents the value of a feature. The blue lines indicate that there is no solar flare in the next 48 hours, and the yellow lines are the opposite. Credit: Space: Science & Technology

Solar flares are solar storm events driven by magnetic field in the solar activity area. When this flare radiation comes to the Earth’s vicinity, the photo-ionization increases the electron density in the D-layer of the ionosphere, causing absorption of high-frequency radio communication, scintillation of satellite communication, and enhanced background noise interference with radar.

Statistics and experience show that the larger the flare, the more likely it is to be accompanied by other solar outbursts such as a solar proton event, and the more severe the effects on the Earth, thus affecting spaceflight, communication, navigation, power transmission and other technological systems.

Providing forecast information on the likelihood and intensity of flare outbreaks is an important element at the beginning of operational space weather forecasting. The modeling study of solar flare forecasting is a necessary part of accurate flare forecasting and has important application value. In a research paper recently published in Space: Science & Technology, Hong Chen from College of Science, Huazhong Agricultural University, combined the k-means clustering algorithm and several CNN models to build a warning system that can predict whether a solar flare would occur in the next 48 hours.

First, the author introduced the data used in the paper and analyzed them from the statistical point of view to provide a basis for the design of the solar flare warning system. To reduce the projection effect, the center of the active region located within ±30°of the solar disk center was selected. After that, the author labeled the data according to the solar flare data provided by NOAA, including the start and end times of the flares, the number of the active region, the magnitude of the flares, etc.

There was a serious imbalance between the number of positive and negative samples in the dataset. To alleviate the imbalance of positive and negative samples, a principle was found to select the events which have positive samples as much as possible. The author visualized the probability density distribution of each feature in all negative samples and all positive samples. It could be easily found that the probability density distributions of the negative samples were all negatively skewed distributions and the characteristics of positive samples were generally larger than those of negative samples. Thus, it was possible to filter out events with positive samples by the feature values of each event.

Afterwards, the author built the whole pipeline with a method containing the following two steps: data preprocessing and model training. To conduct data preprocessing, K-means, an unsupervised clustering method, was used to cluster events to decrease events that only include negative samples as much as possible.

After k-means clustering, all events were divided into three categories, namely category A, category B, and category C. The author found that the ratio of positive samples in category C is 0.340633 which is much larger than that of the whole dataset. Therefore, only the data in category C were chosen as input data on the next stage of algorithm.

In the 2nd stage, the neural networks the author used were Resnet18, Resnet34 and Xception, which are commonly used in deep learning. Three-fourths of samples in category C were randomly chosen. In each event were training data for the neural network models and the rest of the samples were regarded as validation data in the process of training model.

To avoid the influence of dimension, the author also standardized the original data. The standardization method was different from those commonly used. According to the standardization calculation formula, if the label of a sample was predicted to be 1 by the neural network, this sample was regarded as a signal of solar flare which would occur in the next 48 hours. But if it is predicted to be 0, the probability of occurring solar flare in the next 48 hours would be so small that could be ignored.

Then, the author conducted experiments and discussed the results. The author first gave an introduction of experimental setting and then conducted several ablation experiments and comparisons with different models to verify the improvement of k-means clustering algorithm and boosting strategy. Besides, the author also made comparisons between the method used in the experiment and other 13 binary classification algorithms commonly used to present its prediction performance.

The experimental results showed that the prediction performance of the model which integrated several neural networks was better than the one of a single convolutional neural network. Finally, the prediction results of Resnet18, Resnet34, and Xception were combined by boosting strategy. For all networks, recall may be unchanged or even reduced greatly after clustering. However, precision was bound to increase significantly.

After clustering, although the positive sample rate would be greatly improved, from 5% to 34%, nearly 40% of the information of positive samples would also be lost. The author thought this was the main reason why recall remained unchanged or even decreased. It also meant that the number of positive samples predicted in the experiment was less than the one without clustering, but the probability that a predicted positive sample was a true positive was higher.

In contrast with the phenomenon that the prediction performance of other binary classification methods was decreasing or even very poor after clustering, the performance of the author’s method improved by more than 9% after clustering. In conclusion, the two-stage solar flare early warning system consisted of an unsupervised clustering algorithm (k-means) and several CNN models, where the former was to increase the positive sample rate, and the latter integrated the prediction results of the CNN models to improve the prediction performance.

The results of the experiment proved the effectiveness of the method. More information: Jun Chen et al, Two-Stage Solar Flare Forecasting Based on Convolutional Neural Networks, Space: Science & Technology (2022). DOI: 10.34133/2022/9761567 Provided by Beijing Institute of Technology Press Co., Ltd Citation: How a scientist established a two-stage solar flare early warning system (2022, August 19) retrieved 19 August 2022 from https://phys.org/news/2022-08-scientist-two-stage-solar-flare-early.html This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.

Keyword: How a scientist established a two-stage solar flare early warning system

TECH'S NEWS RELATED

Tonga is home to 170 islands. A new one just formed from an underwater volcanic eruption

Nuku’alofa, Tonga. Credit: Unsplash/CC0 Public Domain The Pacific nation of Tonga is made up of 170 islands, but it just welcomed its newest addition—thanks to an underwater volcano. Near the center of the nation’s island formation lies the Home Reef volcano in the South Pacific. On Sept. 10, the ...

View more: Tonga is home to 170 islands. A new one just formed from an underwater volcanic eruption

Potential first traces of the universe's earliest stars

Massive, Population III Star in the Early Universe. This artist’s impression shows a field of Population III stars as they would have appeared a mere 100 million years after the Big Bang. Astronomers may have discovered the first signs of their ancient chemical remains in the clouds surrounding one ...

View more: Potential first traces of the universe's earliest stars

When dangerous toxins teach fundamental biology

Graphical abstract. Credit: Developmental Cell (2022). DOI: 10.1016/j.devcel.2022.09.004 “What our work shows is how a complex in the center of the cell, the ER-Golgi interaction region, controls plasma membrane cholesterol, which is essential for many cellular functions, if not essential for multicellular life,” says Professor Gisou van der Goot ...

View more: When dangerous toxins teach fundamental biology

Every new device Amazon announced at its fall 2022 event

Amazon’s 2022 fall event hardware roundup Kindle Scribe Halo Rise Updated Echo lineup Ring Spotlight Cam Pro and Plus Blink Wired Floodlight Camera Blink Mini Pan Tilt eero PoE 6 and eero PoE Gateway Fire TV Cube Alexa Voice Remote Pro Fire TV Omni QLED Series Amazon hosted its ...

View more: Every new device Amazon announced at its fall 2022 event

Breaking through the mucus barrier

A capsule that tunnels through mucus in the GI tract could be used to orally administer large protein drugs such as insulin.

View more: Breaking through the mucus barrier

Holiday Deals 2022: Here's Where to Get the Best Tech and Gaming Gifts from Sony, Nintendo, and More

Where to Find the Best Deals Best Buy As the holiday season approaches, knowing where to find the best deals on tech and gaming products is critical. Whether it’s to fulfill your personal Christmas wishlist or to give as gifts to loved ones, you don’t want your budget to ...

View more: Holiday Deals 2022: Here's Where to Get the Best Tech and Gaming Gifts from Sony, Nintendo, and More

Upgrade your home office with 74% off this refurbished MacBook Air deal

If you keep up with tech, it’s all too easy to become obsessed with the next big thing, whether that’s the newest iPhone or a tricked-out laptop. All the upgrades are great, but the basic things we need out of our devices haven’t changed much in the last decade ...

View more: Upgrade your home office with 74% off this refurbished MacBook Air deal

Auth0 warns that some source code repos may have been stolen

Authentication service provider and Okta subsidiary Auth0 has disclosed what it calls a “security event” involving some of its code repositories. Auth0’s authentication platform is used to authenticate over 42 million logins each day by more than 2,000 enterprise customers from 30 countries, including the likes of AMD, Siemens, ...

View more: Auth0 warns that some source code repos may have been stolen

Hurricane Ian Postpones Launch of NASA’s SpaceX Crew-5 Mission; New Target Date on Oct. 4

Dogs can smell when we're stressed, study suggests

Brazilian soybean growers' use of biofertilizer examined

Amazon's self-driving units coming 'sooner than people expect'

Lunar glass shows moon asteroid impacts mirrored on Earth

Amazon Kindle Scribe Unveiled: This Jumbo E-Reader Comes With a Stylus

Amazon Fire TV Cube: All the New 3rd-Gen Tricks, from 4K Upscaling to Voice Control

Dall-E Opens Its AI Art Creation Tool to Everyone

Multiple-doped hierarchical porous carbons for superior zinc ion storage

Scientists have a bone to pick with paleontology's portrayal in video games

Scalable and fully coupled quantum-inspired processor solves optimization problems

LHCf continues to investigate cosmic rays

OTHER TECH NEWS

;