trigram
{{short description|Special case of the n-gram, where n is 3}}
{{otheruses}}
{{Refimprove|date=December 2009}}
Trigrams are a special case of the n-gram, where n is 3. They are often used in natural language processing for performing statistical analysis of texts and in cryptography for control and use of ciphers and codes. See results of analysis of "[https://www3.nd.edu/~busiforc/handouts/cryptography/Letter%20Frequencies.html Letter Frequencies in the English Language]".
Frequency
Context is very important, varying analysis rankings and percentages are easily derived by drawing from different sample sizes, different authors; or different document types: poetry, science-fiction, technology documentation; and writing levels: stories for children versus adults, military orders, and recipes.
Typical cryptanalytic frequency analysis finds that the 16 most common character-level trigrams in English are:{{cite book |last= Lewand |first= Robert |title= Cryptological Mathematics |publisher= The Mathematical Association of America |year= 2000 |page= 37 |url= {{google books |id= dx8zM-VeKI8C |page= 37 |text= Most Common Trigraphs in the English Language |plainurl= yes}} |isbn= 978-0-88385-719-9}}{{cite web |url= http://pages.central.edu/emp/LintonT/classes/spring01/cryptography/letterfreq.html |title= Relative Frequencies of Letters in General English Plain text |website= Central College |first= Tom |last= Linton |url-status= dead |archive-date= January 22, 2007 |archive-url= https://web.archive.org/web/20070122235914/http://pages.central.edu/emp/LintonT/classes/spring01/cryptography/letterfreq.html |date= 2001 |series= Cryptography |edition= Spring }}
| class="wikitable sortable" | ||
| Rank | Trigram | Frequency{{cite web |url= http://practicalcryptography.com/cryptanalysis/letter-frequencies-various-languages/english-letter-frequencies/ |title= English Letter Frequencies |website= Practical Cryptography }} (Different source) |
|---|---|---|
| 1 | the | 1.81% |
| 2 | and | 0.73% |
| 3 | tha | 0.33% |
| 4 | ent | 0.42% |
| 5 | ing | 0.72% |
| 6 | ion | 0.42% |
| 7 | tio | 0.31% |
| 8 | for | 0.34% |
| 9 | nde | |
| 10 | has | |
| 11 | nce | |
| 12 | edt | |
| 13 | tis | |
| 14 | oft | 0.22% |
| 15 | sth | 0.21% |
| 16 | men |
Because encrypted messages sent by telegraph often omit punctuation and spaces, cryptographic frequency analysis of such messages includes trigrams that straddle word boundaries. This causes trigrams such as "edt" to occur frequently, even though it may never occur in any one word of those messages.{{cite web |url= https://fuelonline.com/voice-search-seo-voice-seo/ |title= Voice Search SEO |website= Fuelonline }}
Examples
The sentence "the quick red fox jumps over the lazy brown dog" has the following word-level trigrams:
the quick red
quick red fox
red fox jumps
fox jumps over
jumps over the
over the lazy
the lazy brown
lazy brown dog
And the word-level trigram "the quick red" has the following character-level trigrams (where an underscore "_" marks a space):
the
he_
e_q
_qu
qui
uic
ick
ck_
k_r
_re
red
References
{{Reflist}}
{{Natural Language Processing}}
Category:Natural language processing