Analysis of The Efficiency of Several Short-Term Solar Flare Forecasting Techniques Based on Observations of Different Solar Atmospheric Layers

The operational solar flare forecast is an important task in solar physics. It is known to be inherently difficult, and its accuracy remains limited. The forecast quality metrics reported vary significantly, in particular, in recent studies that employ modern machine learning techniques. These studies often report high performance scores; however, they generally lack validation under real-time forecasting conditions, making it difficult to assess their actual effectiveness. Our study presents a comparative analysis of the real-world performance forecasts of solar flares of class ≥C and ≥M that occurred in the period from 2009 to 2024. We compare the forecasts published by the Space Weather Prediction Center on SolarMonitor with the empirical forecasting criteria based on solar radio observations developed by the Northwest Branch of the Special Astrophysical Observatory (SAO) of the Russian Academy of Sciences.