Main Authors
From SONIVIS:Wiki
Main Authors
|
Objective
An article is produced by different authors. This metric helps the analyst to estimate the number of very active users, which work on this specific article. An threshold has to be defined, as from an author is an heavy user.
Explanation
The text of an article is divided into overlapped word sequences with a constant length. Then it has to be revealed in which version the sequence appears at the first time. New added words are related exclusively to the according author.
This procedure is easier to understand with an example: There is an article with three version A,B, C. These article versions are created by Alice, Bob, and Carol.
- Version A: "Iron is an element"
- Version B: "Iron is a chemical element with the symbol Fe in the periodic table"
- Version C: "In the periodic table of the elements iron is a chemical element with the symbol Fe."
If a sequence length of three words is defined, there are two constant word orders "Iron is an element" (version A to C) and "with the symbol Fe" (version B to C) and these word are assigned to Alice and Bob respectively. The word orders "periodic table" (version B and C) is two short to be identified. Therefore the words "In the periodic table of the elements" is assigned to Carol. "a chemical element" exists in version B and C and is from Bob despite the word "a" which is already part of the first sequence from Alice.
For all parts of the article the algorithm calculates the following authors:
| "In the periodic table of the elements | iron is a | chemical element with the symbol Fe." |
| Carol | Alice | Bob |
The in the final text supplemented words are 6 words by Bob (37,5%), 7 words by Carol (43,75%) and 3 words by Alice (18,75%).
Calculation
Specific threshold by which an author is one of the main information creators.
Reference
- Definition from Jah
- Example (in German)
- Textual description adapted and translated from Jacob Voss (pages 52-55) (in German)
- Similar approach in A Method for Measuring Co-authorship Relationships in MediaWiki (at WikiSym 2008)
