The Weber (1970) and Van Orden (1987) Algorithm for determining spelling
similarity
Note: Your browser needs to be Java-enabled.
Type in two words in the textboxes on the leftside of window. The box
on the right will provide you with the measure of Graphic Similarity, as
described in Van Orden (1987, p. 196)
This measure is based on:
- F: number of pairs of adjecent letters in the same order, shared
by pairs
- V: number of pairs of adjecent letters in reverse order, shared by
pairs
- C: number of single letters shared by word pairs. Note: for meet
/ meet this is four!!
- A: average number of letters in the two pairs
- T: ratio of shorter word to longer word
- B: if first two letters are the same B = 1 else B = 0
- E: if last two letters are the same E = 1 else E = 0
Then the Graphic Similarity = 10([(50F + 30V + 10C)/A] + 5T + 27B + 18E)
Note: Van Orden (1987) calculates Orthographic Similarity, by determining
the ratio between GS of word one with itself and GS of word 1 and word 2.
Note: Use this at your own risk! I don't guarentee this program is free
of bugs
Note: Marc Brysbaert, Sarah Haywood, and Wouter Duyck pointed out
a user interface problem with this applet! When you change the value of word
1 you must type <Enter>, otherwise the comparison will be done
with the default word (meet).
To determine spelling similarity, type in a word
followed by <ENTER> in each text box