Objective method for forecasting solar flares. Environmental research papers

Solar parameters derived from the region analysis program at the NOAA Space Environment Services Center (SESC) are submitted to a multivariate discriminant analysis (MVDA) in which the parameters relevant to flare prediction are identified and incorporated in a classification procedure to produce a flare forecast. The analysis uses two years of data (6095 solar active region-days). The MVDA forecast is compared with a subjective forecast derived from the SESC forecast during the same period, and is found to have greater accuracy overall. Specific recommendations are made concerning the application of the technique in a forecasting operation, and in the types of data required for future improvement.