Comparison of metrics for measuring Wikipedia ecology: characteristics of self-consistent metrics for editor scatteredness and article complexity

• Published in: Artificial life and robotics Vol. 28; no. 1; pp. 62 - 66
• Main Authors: Ogushi, Fumiko, Shimada, Takashi
• Format: Journal Article
• Language: English
• Published: Tokyo Springer Japan 01.02.2023; Springer Nature B.V
• Subjects: Artificial Intelligence; Complexity; Computation by Abstract Devices; Computer Science; Control; Editors; Encyclopedias; Evaluation; Mechatronics; Original Article; Robotics; Wikipedia; Collective knowledge; Complex network; Self-consistent metrics
• ISSN: 1433-5298; 1614-7456
• DOI: 10.1007/s10015-022-00819-x

Wikipedia, an example of a collective knowledge space, improves the quality of its editors and articles in a self-organized manner. To measure the quality of the editors and articles on Wikipedia, self-consistent metrics for the network defined by the edit relationship have been introduced previously. This scatteredness–complexity measure can evaluate the editors and articles more sensitively than the local characteristics such as degrees of the network and capture well the editors’ activity and the articles’ level of complexity. Here, we show that the scatteredness–complexity measure are equivalent to their relative degrees in a random network, where the editors select articles randomly for editing and the editors and the articles are uncorrelated. In addition, the distributions of the editor scatteredness and article complexity become smoother when the network is randomized and loses its detailed local structure eliminating the correlation between the editors and articles. When the degree distributions of the editors or articles are changed and become uniform in the randomized network, the distributions of the editor scatteredness or article complexity become flatter, respectively. This results suggest that the scatteredness–complexity measure reflects not only the degree distribution of the editors or articles but also the local network structure. Analysis using our self-consistent metrics capturing the local structure would be useful for evaluating collective knowledge.