$RMSAE$ vs actual-DOA for the (a) SVR, (b) RR, (c) PR1, (d) PR2, (e) DAS, and (f) MUSIC algorithm. Training of regression model is done at SNR = 26dB and testing is performed at SNR = 18dB