Solar flares are the result of sudden discharge of the magnetic energy produced in the sun's atmosphere, which causes a flash observation on its surface. The energy released is equal to numerous nuclear explosions, enough to damage communication and power grids on the Earth. Therefore, an early prediction is required to avoid the havoc caused by solar flares.
Using artificial intelligence, researchers have found a solution to predict the flares by analyzing a huge set of data produced by the Solar Dynamics Observatory. The study mainly focuses on the strength and direction of the flares, also called as the vector magnetic field, but only for a particular part of the sun. The researchers faced difficulty in anticipating the solar flares before they came across the machine learning technique of the artificial intelligence. The machine learning software would establish set of information categories by observing different patterns and comparing the relevancy with a category.
How the system work?
The analysts recorded regions producing flares and not from the database containing more than 2,000 active regions. They categorized the regions into 25 different features such as field and current gradient, energy, etc. To train the learning machine, only 70 percent data were fed to it and the rest 30 percent is provided for the analyzing; to test the accuracy of prediction of the flares. Using only a few features out of 25, the machine provided the result about the region that would flare and the area that will not.
In spite of the fact that others have utilized distinctive systems to think of comparative results, machine learning gives a noteworthy change on the grounds that automated examination is quicker and could give prior warnings of sunlight based flares. However, the data provided was taken from the sun's surface which similar to predicting weather on Earth by noting surface temperature. The researchers are now trying the same method using the data from the solar atmosphere. This experiment shows the way by which we can predict other natural calamities like earthquake, tsunami and can prevent the disaster before it happens. The study is a ray of hope for preventing the disasters that caused humans to suffer from years.
Copyright © 2013 A.I.M.S. Technology Pvt. Ltd.. All Rights Reserved.