Edit Categories and Editor Role Identification in Wikipedia

Edit Categories and Editor Role Identification in Wikipedia Diyi Yang♣, Aaron Halfaker♦, Robert Kraut♠, Eduard Hovy♣ ♣ Language Technologies Institute, Carnegie Mellon University {diyi,hovy}@cmu.edu ♦ Wikimedia Foundation ahalfaker@wikimedia.org ♠ Human-Computer Interaction Institute, Carnegie Mellon University robert.kraut@cmu.edu Abstract In this work, we introduced a corpus for categorizing edit types in Wikipedia. This fine-grained taxonomy of edit types enables us to differentiate editing actions and find editor roles in Wikipedia based on their low-level edit types. To do this, we first created an annotated corpus based on 1,996 edits obtained from 953 article revisions and built machine-learning models to automatically identify the edit categories associated with edits. Building on this automated measurement of edit types, we then applied a graphical model analogous to Latent Dirichlet Allocation to uncover the latent roles in editors’ edit histories. Applying this technique revealed eight different roles editors play, such as Social Networker, Substantive Expert, etc. Keywords: Edit Category, Role Identification, Topic Modeling 1. Introduction Distributed work teams in online communities have be- come increasingly important in creating innovative prod- ucts, such as GNU, Linux and Wikipedia. Millions of vol- unteers participate in the online production communities, exchange their expertise and ideas, and collaborate to pro- duce complex artifacts. For example, in Wikipedia, edi- tors take up different responsibilities, when editing articles, based on their interest and expertise. Some, for example, might add substantive new content to articles while others may focus on copy-editing. Better understanding of the par- ticipants and what they do, how they behave can make these online communities more successful. Thus, the problem is, given two revisions of a piece of text, what actions have been done by a user to transform the original version into the new one, and who that user is? Our goal in this work is to develop fine-grained taxonomy to categorize editors’ el- emental actions when editing Wikipedia articles and design new methods to identify roles that editors exhibit. The problem of identifying editors’ roles in Wikipedia has attracted significant attention (Arazy et al., 2015; Ferschke et al., 2015). Numerous studies have discussed how to identify roles based on users’ behavioral regularities and social network signatures (Welser et al., 2011; Welser et al., 2007). Most research classifies editors based either on their edits in different namespaces or via the user attributes such as access privileges (Arazy et al., 2015), barnstars (Kriplean et al., 2008), etc. Classification based on users’ attributes is relatively accurate, but this information is not available for many active editors and is insufficient in ex- plaining the nature of an editor’s work. While classification based on edit histories can be constructed for most active editors, current approaches focus on simple edit counts and access privileges fail to provide a finer grained description of the work actually performed in an edit. For example, it cannot tell the difference between an editor who copy-edits a paragraph and an editor who inserts markup or template to an article. In this work, we extend Daxenberger’s fine grained taxon- omy …

Content truncated to 500 words