Authors: Anastasia Kuske (NJIT), Bin Chen (NJIT), Gelu Nita (NJIT)

Type III radio bursts, which are characterized by rapid frequency drifts, are frequently observed solar radio phenomena. Type III bursts are believed to be generated by streams of fast electron beams traveling through the solar corona; they can occur as isolated events or as a group, sometimes referred to as type III storms, and are indicative of the highly dynamic processes within the corona. However, it remains an open question how these bursts are precisely linked to different types of solar activity and the specific mechanisms driving their variability, largely due to reliance on spacecraft data and a lack of imaging of these events. Owens Valley Radio Observatory’s Long Wavelength Array (OVRO-LWA) is a newly commissioned ground-based radio facility that provides continuous spectral imaging of the Sun in the metric wavelengths (15-88 MHz). Thanks to its extremely high sensitivity, OVRO-LWA has observed many instances of these bursts over more than six months to date. They serve as an excellent data source for exploring their source region, statistics, and long-term evolution. We present an advanced machine learning-based approach for identifying type III solar radio bursts using OVRO-LWA data. The identified burst periods are then used for obtaining temporally and spectrally resolved radio images to obtain the type III burst source regions. We discuss our findings on Type III burst identification, highlighting the effectiveness of our random forest classifier in differentiating bursts from the background, and the success of our segmentation method in accurately identifying and analyzing bursts.