Intro to Predictive Coding: Overview & Interpretation of Terminology June 2014 | Page 14

If the answer to at least one of these five questions is yes, then
there is one more question to consider.

• Does your collection contain more than about 5,000 text
documents?
Predictive coding does not require a large set of documents,
but it’s value tends to grow disproportionately as the size of the
document collection grows, because the effort typically
required to train a system does not grow or does not grow as
quickly as the size of the document collection increases. Small
collections can require almost the same level of training effort
as large collections do.