Integration of multiple network views in Wikipedia

Abstract

One of the challenges in network data analysis is the determination of the most informative perspective on the network to use in analysis. This is particularly an issue when the network is dynamic and is defined by events that occur over time. We present an example of such a scenario in the analysis of edit networks in Wikipedia -the networks of editors interacting on Wikipedia pages. We propose the prediction of article quality as a task that allows us to quantify the informativeness of alternative network views. We present three fundamentally different views on the data that attempt to capture structural and temporal aspects of the edit networks. We demonstrate that each view captures information that is unique to that view and propose a strategy for integrating the different sources of information.

Understanding the Signature of Controversial Wikipedia Articles through Motifs in Editor Revision Networks

“…Controversial articles have become an increasing point of focus, and characterised as such by Wikipedia. Automated methods for classifying articles have received much attention (e.g., [21,27,29]. The associated revision log for Wikipedia articles has been shown to provide a basis to examine potential controversy through examining the collaborative behaviour of individual editors within an article [23] or across multiple articles [28].…”

“…The associated revision log for Wikipedia articles has been shown to provide a basis to examine potential controversy through examining the collaborative behaviour of individual editors within an article [23] or across multiple articles [28]. An article's revision log identifies the structure underlying temporal interactions [27], and provides insight into how articles and contributors' habits may evolve over time [9]. Features from the aggregation of this, such as number of edits, revision, and previous version restorations have been shown to correlate (e.g., [23]).…”

“…For each article, we construct a revision network (in the same manner as [11,27]) where nodes represent unique editors and directed edges show that an editor added a revision after another editor. We traverse the revision log list to build a network that spans the article's lifetime, adding nodes and edges as they appear in each event.…”

Frequent Subgraph Mining Based Collaboration Pattern Analysis for Wikipedia

“…Furthermore, based on various networks, applying the analysis of motifs to study the properties of networks is a good way to research collaboration [14]. Wu and Cunningham [33] integrated three network views to predict the qualities of articles, among which are the temporal network that considers article revision histories as sequence of editor's interactions, ego network which is a static bipartite network with edges representing edits performed by editors to articles and trajectory network which is representative of revisions of an article through a sequence of editors. Then they conducted a binary classification task, based on motifs discovered in the three networks, to assess the effectiveness of their method.…”

“…Due to many other researchers [5,33,34], applying various classification algorithms, classified articles into two classes, we also utilized the combination of 11 features presented in Figure 12.b, based on six types of classification algorithms, to classify articles into two classes, respectively. From Figure 12.b, we can see that the classification accuracy of Ensemble Learning (ensembles for Boosting, Bagging and Random Subspace) is the highest, which is nearly 100%.…”