Table 2 Setup of hyperparameters in gplearn toolkit for GPSR.

From: Exploring high thermal conductivity polymers via interpretable machine learning with physical descriptors

Parameter

Value

Generations

300

Population size in every generation

5000

Probability of crossover (pc)

[0.30,0.90], step = 0.05

Probability of subtree mutation (ps)

[(1-pc)/3,(1-pc)/2] (step = 0.01)

Probability of hoist mutation (ph)

[(1-pc)/3,(1-pc)/2] (step = 0.01)

Probability of point mutation (pp)

1-pc-ps-ph

Function set

\(\{+,-,\times ,\div,\sqrt{x},\mathrm{ln}x,\left|x\right|,-x,1/x\)}

Parsimony coefficient

0.001, 0.003, 0.005

Metric

R2

Stopping criteria

0.900

Random_state

0, 1, 2, 3, 4

Init_depth

[2, 6], [4, 8], [6, 10], [2, 10]