Skip to main content

Table 3 The RL algorithm achieves a faster running time when compared to Super.Complex

From: Molecular complex detection in protein interaction networks through reinforcement learning

Method

Processor specifications

Training

Prediction

Post-processing

Total time (s)

No. of cores

Time (s)

No. of cores

Time (s)

No. of cores

Time (s)

RL Algorithm

M1 chip @ 3.2 GHz

1

9

8

320

1

11

340

Super.Complex

Intel(R) Xeon(R) E5-2699 v3 @ 2.30 GHz

20

540

72

17

1

112

669

  1. The time reported here corresponds to runs using the best overlap threshold found for both methods and using the best heuristic found in the case of Super.Complex