Average word length |
---|
8.3943 |
word length | percentage |
---|---|
1 | 0.2631 |
2 | 1.0074 |
3 | 2.7681 |
4 | 6.1489 |
5 | 8.6749 |
6 | 11.2277 |
7 | 13.1558 |
8 | 11.7030 |
9 | 11.6342 |
10 | 9.4132 |
11 | 7.4702 |
12 | 5.6587 |
13 | 3.9399 |
14 | 2.7412 |
15 | 1.7487 |
16 | 1.0791 |
17 | 0.6696 |
18 | 0.3408 |
19 | 0.1883 |
20 | 0.1046 |
21 | 0.0568 |
22 | 0.0239 |
23 | 0.0149 |
24 | 0.0209 |
25 | 0.0060 |
26 | 0.0030 |
27 | 0.0030 |
29 | 0.0030 |
In this subsection we ignore the fact that words have different frequencies. So for the average word length, each word is considered equally. For a fixed word length, we count the number of different words having this length.
The plot of the word length against the number of words of this length usually has a clear maximum between 10 and 15. Moreover, with a logarithmic scale of the y-axis, we get a nearly linear part between length 15 and 40.
Average word length is one of the classic parameters for a language.
Counting without multiplicity makes average word length depending on the corpus size. A larger corpus contains more words, and the additional words are usually longer. Hence, average word length should increase with corpus size.
Average word length:
select avg(char_length(word)) from words where w_id>100;;
Data for large table:
SELECT @all:=count(*) from words where w_id>100;
select char_length(word), 100*count(*)/@all from words where w_id>100 group by char_length;
Do we have the linear part between 15 and 40 for (nearly) all languages?
Where does it come from?
Calculate and compare the slope!
3.5.1.2 Words by Length with multiplicity