N-Gram Extractor

The N-Gram routine formats an input text as strings of a chosen length (words 1, 2, 3, 4; words 2, 3, 4, 5; etc) and counts up those that are repeated, whether independently meaningful ('nice day', 'good looking,' etc) or not ('of the,' 'in a,' etc). Repeated sequences are also known as lexical bundles. Note that the A and B modes in this routine are quite different : A is an Demo of the key ideas but B is more powerful (if large output disappears off right screen, scroll with arrow keys) .

Phrase Extractor

Phrase Extractor is N-Gram with an extra step. It tries to reduce the N-Gram list down to just meaningful phrases (collocations) by calculating the mutual information (MI) provided by the individual words of the phrase. This is the ratio of the frequency of a phrase to the frequencies of its individual words (see examples). MI needs sizeable or specialized corpus for interesting result.

Phrase Profiler

Phrase Profiler finds a list of phrases in a text or corpus, for example a list of sentence transitions ('On the other hand') or academic collocations ('In this study')