In addition, natural languages are
highly ambiguous, as two languages seldom express the same
content in the same way [12]. Although hybrid approaches also
exist, MT systems can be broadly classified into two main
categories, corpus-based and rule-based, according to the nature
of the linguistic knowledge being used. The rule-based MT
systems use knowledge in the form of rules, explicitly coded by
human experts, which attempt to codify the translation process.
Instead, corpus-based MT systems use large collections of
parallel texts (i.e. pairs consisting of a text in a source language
and its translation into a target language) as the source of
knowledge from which the engine learns how to perform
translations