
178 Applying Analytics: A Practical Introduction
a match ca n be missed or a do c ument that should not have been a match can
be mistakenly included. These are known, respectively, as errors of omission
and commission. These type s of errors are not limited to text mining; analo -
gous behavior in the image of the flock of birds is ea sy to imagine. You may
miss a bird because your view is partially blocked by one in front of it (omis-
sion), or your software may mistake a jet plane in the background as a bird
(commission).
Clearly, building a piece of software to extract metadata from unstructured
data is fraught with subtlety and complexity. It also ...