State Evaluation Method of Distribution Terminal Based on Deep Reinforcement Learning

Table 5

Prediction performance of the proposed network model.

Learning rate	Training parameters	ER
0.050	10000	298.01
0.100	10000	286.32
0.150	10000	240.95
0.050	20000	279.19
0.100	20000	257.03
0.150	20000	232.99