word rank | frequency | word |
---|---|---|
1 | 1011 | като |
2 | 931 | през |
3 | 679 | През |
4 | 450 | това |
5 | 369 | след |
word rank | frequency | word |
---|---|---|
1 | 701 | година |
2 | 289 | години |
3 | 172 | селото |
4 | 140 | където |
5 | 132 | всички |
word rank | frequency | word |
---|---|---|
1 | 179 | България |
2 | 105 | началото |
3 | 86 | страната |
4 | 85 | получава |
5 | 82 | различни |
word rank | frequency | word |
---|---|---|
1 | 59 | обикновено |
2 | 59 | продължава |
3 | 35 | останалите |
4 | 29 | последните |
5 | 27 | значително |
word rank | frequency | word |
---|---|---|
1 | 31 | представлява |
2 | 29 | Първоначално |
3 | 29 | правителство |
4 | 25 | изключително |
5 | 24 | управлението |
word rank | frequency | word |
---|---|---|
1 | 23 | Великобритания |
2 | 18 | правителството |
3 | 16 | Освобождението |
4 | 16 | производството |
5 | 14 | Министерството |
Slope |
---|
-0.9854058054362589 |
Slope |
---|
-0.752574989159953 |
Slope |
---|
-0.6235773074405634 |
Slope |
---|
-0.5484550065040282 |
Slope |
---|
-0.6020599913279624 |
Slope |
---|
null |
The validity of Zipf’s law for all words does not imply its validity for words of fixed length. The tables show the most frequent words for fixed word lengths 4, 6, 8, …, 14.
The diagram implies that Zipf’s law is still valid, at least for larger word length.
The diagrams are interesting for language comparison. In the case of a diagram of very different shape there might have been problems in preprocessing which should be visible elsewhere. Or, more interesting, there might be special features inherent to this language.
For length 8:
SELECT freq ,word FROM words WHERE w_id>100 and char_length(word)=8;
Explain why we see better agreement with Zipf’s law for larger word length.
Zipf’s law (Classic version)