Data Description
Predict the number of common solar flares.
The database contains 3 classes, one for the number of times a certain type of solar flare occurred in a 24 hour period.
Each instance represents captured features for 1 active region on the sun.
Attribute Description
1. *class* - code for class (modified Zurich class) (A,B,C,D,E,F,H)
2. *largest_spot_size* - Code for largest spot size (X,R,S,A,H,K)
3. *spot_distribution* - Code for spot distribution (X,O,I,C)
4. *activity* - activity (1 = reduced, 2 = unchanged)
5. *evolution* - evolution (1 = decay, 2 = no growth, 3 = growth)
6. *previous_activity* - previous 24 hour flare activity code (1 = nothing as big as an M1, 2 = one M1, 3 = more activity than one M1)
7. *complex* - historically-complex (1 = Yes, 2 = No)
8. *complex_path* - whether region become historically complex on this pass across the sun's disk (1 = yes, 2 = no)
9. *area* - area (1 = small, 2 = large)
10. *area_largest* - area of the largest spot
11. *c_class_flares* - number of C-class flares production by this region in the following 24 hours (common flares), target feature
12. *m_class_flares* - number of M-class flares production by this region in the following 24 hours (moderate flares)
13. *x_class_flares* - number of X-class flares production by this region in the following 24 hours (severe flares)