Table 5 A comparison between predictions by DFT and ML of ‘dominating impurities’ in CdTe.

From: Machine-learned impurity level prediction for semiconductors: the example of Cd-based chalcogenides

Verdict

Cd-rich

Te-rich

 

Predicted

\(\%\) of total

Predicted

\(\%\) of total

False positives

5

1.59

3

0.95

False negatives

10

3.17

6

1.90

True negatives

272

86.35

275

87.30

True positives

28

8.89

31

9.84

  1. True positives refer to the cases that were predicted to be dominating by both DFT and ML, and true negatives are the cases predicted to be non-dominating by both. False positives were predicted to be dominating by only ML whereas false negatives were predicted to be dominating by only DFT.
Back to article page