Automatic Reputation Assessment in Wikipedia

Abstract

The online encyclopedia Wikipedia is predominantly created by anonymous or pseudonymous authors whose knowledge and motivations are unknown. For that reason there is an uncertainty in terms of their contribution quality. An approach to this problem is provided by automatic reputation systems, which have been becoming a new research branch in the recent years. In previous research, different metrics for automatic reputation assessment have been suggested. Nevertheless, the metrics are evaluated insufficiently and considered isolated only. As a result, the significance of these metrics is quite unclear. In this paper, we compare and assess seven metrics, both originated from the literature and new suggestions. Additionally, we combine these metrics via a discriminant analysis to deduce a significant reputation function. The analysis reveals that our newly suggested metric editing efficiency is particularly effective. We validate our reputation function by means of an analysis of Wikipedia user groups.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 12 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.