Source-linked AI summary
Enhancing Navigation on Wikipedia with Social Tags
Arkaitz Zubiaga
TL;DR
Wikipedia’s search and navigation can be limited by missing content terms and the structure of its categories and links. The paper proposes a prototype combining Wikipedia articles with tags collected from Delicious. Its reported analyses find many new terms and additional ways to navigate, search, and identify popular articles.
Problem
Wikipedia’s search may miss articles when relevant words are absent from their content, while category and link navigation depend on defined categories and available links.
Source: §1 · p. 2Method
The authors built a prototype applying Delicious tag information to English Wikipedia articles, retaining articles tagged by at least 10 Delicious users.
Source: §4 · p. 6, §4.1 · p. 6Results
The authors report that tag information provides a high number of new terms for articles, adding terms for search and navigation.
Source: §5 · p. 9Takeaways & Limitations
The authors conclude that tags offer new navigation routes, additional search terms, and a way to discover popular articles on Wikipedia.
Source: §5 · p. 9Open-vocabulary tags can have equivalent variants, such as “science fiction” and “sci-fi,” so relations between tags may be needed to avoid tag mess.
Source: §3.3 · p. 5
Abstract
from arXiv · showhide
Social tagging has become an interesting approach to improve search and navigation over the actual Web, since it aggregates the tags added by different users to the same resource in a collaborative way. This way, it results in a list of weighted tags describing its resource. Combined to a classical taxonomic classification system such as that by Wikipedia, social tags can enhance document navigation and search. On the one hand, social tags suggest alternative navigation ways, including pivot-browsing, popularity-driven navigation, and filtering. On the other hand, it provides new metadata, sometimes uncovered by documents' content, that can substantially improve document search. In this work, the inclusion of an interface to add user-defined tags describing Wikipedia articles is proposed, as a way to improve article navigation and retrieval. As a result, a prototype on applying tags over Wikipedia is proposed in order to evaluate its effectiveness.
1. INTRODUCTION
Social tags offer collaborative, weighted metadata that can broaden retrieval and navigation beyond article content. The paper asks whether tags can improve Wikipedia’s existing search, category, and link-based navigation.
- Collaborative tagging aggregates users’ tags into weighted metadata, including open-vocabulary terms that may not appear in a document’s content.Prior work cited in the passage reports that tags improve information retrieval and navigation.Source: §1 · p. 1
- Wikipedia’s keyword search may miss articles when relevant words are absent from their content, while category and link navigation are limited by their defined structure and available links.Source: §1 · p. 2
- The paper investigates whether social tagging can improve article search and retrieval alongside Wikipedia’s existing navigation methods.Source: §1 · p. 2
2. SOCIAL TAGGING
Social tagging makes user-assigned keywords publicly useful as metadata, with collaborative systems combining contributions into weighted tags. Tag clouds summarize popular tags across a site.
- Tags are open-vocabulary keywords that users assign to describe an item and support its retrieval, unlike hierarchical taxonomic categories.Source: §2 · p. 2
- In simple tagging, only a resource’s owner adds tags, whereas collaborative tagging lets all users contribute tags to a resource.Source: §2 · pp. 2–3
- Collaborative contributions produce a weighted set of tags for each resource, and social tagging systems may display popular tags in a tag cloud.In a tag cloud, larger font size indicates that a tag appears in more resources.Source: §2 · p. 3
3. INCLUDING TAGS INTO NAVIGATION
The paper presents tags as a bottom-up complement to Wikipedia’s taxonomy, offering additional navigation routes and search terms while requiring care with open-vocabulary variants.
3. INCLUDING TAGS INTO NAVIGATION
- Pivot-browsing, popularity-driven navigation, and filtering let users explore related tags, find highly tagged documents, or exclude unwanted tags.Source: §3 · p. 4
3.1 Benefits of Tagging
- An open vocabulary lets users add new tags without checking an existing list, and aggregating their tags gives articles weighted descriptors.Source: §3.1 · pp. 4–5
3.2 Enhancing Taxonomies by means of Tagging
- Adding tags to a taxonomy supplies alternative navigation, more article terms for search, and bottom-up classification alongside top-down categories.Source: §3.2 · p. 5
3.3 Avoiding Tag Mess
- Open-vocabulary variants such as “science fiction,” “science-fiction,” and “sci-fi” can fragment tags; relating equivalent tags is proposed to reduce this tag mess.The passage suggests a large community could help organize tag relations.Source: §3.3 · p. 5
4. EXPERIMENTS
The experiments evaluate a Delicious-tagged Wikipedia prototype through navigation and tag-occurrence analyses. The results describe new browsing options and terms that tags add beyond article text and categories.
4.1 Dataset Generation
- The prototype uses Delicious tags collected for English Wikipedia articles from April 2009, retaining articles tagged by at least 10 users and producing a dataset of 20,764 articles.The tags wikipedia, reference, and wiki were removed as irrelevant to describing individual articles.Source: §4.1 · p. 6
4.2 Results
- Qualitative navigation through the prototype suggests that pivot-browsing, filtering, and popularity-based navigation are interesting new ways to explore articles.The prototype was developed to evaluate whether tagging could help Wikipedia navigation and search.Source: §4.2 · p. 6, §4 · p. 6
- Aggregating weighted tags across articles produced a tag cloud with programming, psychology, science, research, software, and architecture standing out.Figure 5 shows the resulting tag cloud for the experiments.Source: §4.2 · p. 6, Figure 5 · p. 7
- Tags appear in less than 50% of article content and less than 8% of categories, indicating many new terms emerge from tags.The analysis also considers tag occurrence across article content, category names, and the whole document.Source: §4.2 · p. 7
- The more tags an article has, the more new terms emerge, as shown by the tag-presence analysis.Figure 6 presents tag-presence percentages according to the number of tags an article has.Source: §4.2 · p. 7, Figure 6 · p. 8
- Examples include programming for Framework, mathematics for Zipf’s law, and audio for List of Internet stations, although those terms are absent from the articles’ text.Source: §4.2 · p. 8
5. CONCLUSIONS
The conclusions report positive experimental results for adding tags to Wikipedia and describe improvements to navigation and search. They distinguish new terms from new tag-based browsing routes and contrast bottom-up tags with Wikipedia’s category hierarchy.
5. CONCLUSIONS
- The prototype’s qualitative analysis identifies pivot-browsing, popularity, and filtering as new ways to navigate through articles.Source: §5 · p. 8
- Quantitative analysis finds a high number of new terms for each article when tag information is used, adding terms for search and navigation.Source: §5 · p. 9
- Tag navigation uses bottom-up classification for same-level cross-navigation, whereas categories require moving up the tree to find new leaf categories.Source: §5 · p. 9
- The experiments show positive results and encourage using a tagging system to improve navigation and search on Wikipedia.Source: §5 · p. 9
- Tags provide new terms, new navigation routes, search-engine assistance, and a way to discover popular articles.Source: §5 · p. 9