Home Software Services About Contact usearch manual
Taxonomy classification errors

See also
 
Taxonomy benchmark home
 
Validating taxonomy classifiers
  Training taxonomy classifiers
  UTAX algorithm
  Splitting a taxonomy reference set
  Defining "accuracy" of a taxonomy classifier

False positive error
Classifier predicts an incorrect taxon for the given level (family, genus etc.).

Misclassification error
Type of false positive error. Classifier predicts a wrong name at the given level (family, genus etc.) when at least one reference sequence for the correct taxon is present in the training set.

Overclassification error
Type of false positive error. Classifier predicts a name at the given level (family, genus etc.) when there are no reference sequences for the correct taxon in the training set. See taxonomy overclassification and underclassification errors for further discussion.

False negative error
Classifier does not predict a name for the given taxon (family, genus etc.) when at least one reference sequence for the correct taxon is in the training set.

Underclassification error
Classifier does not predict a name at the given level (family, genus etc.) when there are reference sequences for the correct taxon in the training set. See taxonomy overclassification and underclassification errors for further discussion. All false negatives are underclassification errors so there is really no need for a new term.

 
Taxon is present in training data Taxon is predicted by classifier Correct name is predicted Correct / Error Result
Yes Yes Yes Correct True positive
Yes Yes No Error False positive
(misclassification)
Yes No - Error False negative (underclassification)
No Yes - Error False positive
(overclassification)
No No - Correct True negative