Abstract

Recent research has shown that different tagging motivation and user behavior can effect the overall usefulness of social tagging systems for certain tasks. In this paper, we provide further evidence for this observation by demonstrating that tagging data obtained from certain types of users – so-called Categorizers – outperforms data from other users on a social classification task. We show that segmenting users based on their tagging behavior has significant impact on the performance of automated classification of tagged data by using (i) tagging data from two different social tagging systems, (ii) a Support Vector Machine as a classification mechanism and (iii) existing classification systems such as the Library of Congress Classification System as ground truth. Our results are relevant for scientists studying pragmatics and semantics of social tagging systems as well as for engineers interested in influencing emerging properties of deployed social tagging systems.

Share/Save

Part of our organizational streamlining involves cutting our investment in underperforming or off-strategy products to put better focus on our core strengths and fund new innovation in the next year and beyond.

We continuously evaluate and prioritize our portfolio of products and services, and do plan to shut down some products in the coming months such as Yahoo! Buzz, our Traffic APIs, and others. We will communicate specific plans when appropriate.

Hopefully, someone will keep Delicious alive as MerchantCircle did with Bloglines. Let’s see how this turns out..

Update: Delicious announced that they’re not closing down the social bookmarking site, and they’ll look for a new home for it outside Yahoo.

Share/Save

Amongst others, Amazon uses the latter approach to suggest tags when nobody has annotated a book yet. In this case, we can see a subsection “Suggested Tags from Similar Products” for that book. Nonetheless, this doesn’t seem to be a good approach, and it sometimes suggests us weird tags, as happens for O’Reilly’s book “Cassandra: The Definitive Guide” [retrieved on November 3, 2010]:

Cassandra: The Definitive Guide on Amazon

As stated by Delicious founder Joshua Schachter:

Automatic tags lose a lot – doesn’t help the user really achieve their goals. That’s why the ‘add to del.icio.us’ badges don’t let you suggest tags.

Well, I think we should pay attention to Schachter’s quote, and let the users tag by themselves, without suggesting them automatic tags.

More posts will come soon on tag suggestions.

Share/Save

These are the terms I have collected so far:

Tagging & Social Tagging

Tagging: Tagging is an open way to assign tags or keywords to resources or items (e.g., web pages, movies or books), in order to describe them. This enables the later retrieval of the resources in an easier way, using tags as resource metadata. As opposed to a classical taxonomy-based categorization system, they are usually non-hierarchical, and the vocabulary is open, so it tends to grow indefinitely. For instance, a user could tag this blog as social-tagging, research and blog, whereas another user could use web2.0, social-bookmarking and tagging tags to annotate it.

Social tagging: A tagging system becomes social when its tag annotations are publicly visible, and so profitable for anyone. The fact of a tagging system being social implies that a user could take advantage of tags defined by others to retrieve a resource.

Tag cloud: In order to enable visual browsing, social tagging tools typically provide an interface model known as tag cloud. When users access the information in these structures, it is presented in the form of a cloud consisting of the most popular tags, where the larger is the font size of a tag, the more popular it is on the site. Typical tag clouds have between 50 and 200 tags, and tag weights are represented using font sizes or other visual clues. In addition, tags can be sorted in alphabetical, size-based or random order, and users can usually customize clouds with different fonts, layouts, and color schemes. These structures are particularly useful for browsing or information discovery, because they provide a visual summary of the content in the collection. More information on Wikipedia: Tag cloud.

Social Bookmarking & Social Cataloging

Social bookmarking: Delicious, StumbleUpon and Diigo, amongst others, are known as social bookmarking sites. They provide a social means to save web pages (or other online resources like images or videos) as bookmarks, in order to retrieve them later on. In contrast to saving bookmarks in user’s local browser, posting them to social bookmarking sites allows the community to discover others’ links and, besides, to access the bookmarks from any computer to the user itself. In these systems, bookmarks represent references to web resources, and do not attach a copy of them, but just a link. Note that social bookmarking sites do not always rely on social tags to organize resources, e.g., Reddit is a social bookmarking approach to add comments on web pages instead of tags. The use of social tags in social bookmarking systems is a common approach, though. For more information, see: Social bookmarking and List of social bookmarking sites on Wikipedia.

Social cataloging: They are quite similar to social bookmarking sites in that resources are socially shared but, in this case, offline resources like music, books or movies are saved. For instance, LibraryThing allows to save the books you like, Hulu does it for movies and TV series, and Last.fm for music-related resources. As in social bookmarking sites, tags are the most common way to annotate resources in social cataloging sites.

Folksonomy & Personomy

Folksonomy: As a result of a community tagging resources, the collection of tags defined by them creates a tag-based organization, so-called folksonomy. A folksonomy is also known as a community-based taxonomy, where the classification scheme is plain, there are no predefined tags, and therefore users can freely choose new words as tags. A folksonomy is basically known as weighted set of tags, and may refer to a whole collection/site, a resource or a user. A summary of a folksonomy is usually presented in the form of a tag cloud.

Personomy: Personomy is a neologism created from the term folksonomy, and it refers to the weighted set of tags of a single user/person. It summarizes the topics a user tags about.

Simple Tagging & Collaborative Tagging

To the best of my knowledge, these two terms were first coined by Gene Smith, in the book Tagging: People-Powered Metadata for the Social Web. Previously, Thomas Vander Wal referred to them as Narrow Folksonomy and Broad Folksonomy.

Simple tagging: users describe their own resources or items, such as photos on Flickr, news on Digg or videos on Youtube, but nobody else tags another user’s resources. Usually, the author of the resource is who tags it. This means no more than one user tags an item. In many cases, like in Flickr and Youtube, simple tagging systems include an attachment to the resource, and not just a reference to it.

Collaborative tagging: many users tag the same item, and every person can tag it with their own tags in their own vocabulary. The collection of tags assigned by a single user creates a smaller folksonomy, also known as personomy. As a result, several users tend to post the same item. For instance, CiteULike, LibraryThing and Delicious are based on collaborative tagging, where each resource (papers, books and URLs, respectively) could be tagged (therefore annotated) by all the users who considered it interesting.

For more information on simple and collaborative tagging, see my previous post: What are social tags?

Share/Save

Have a look at it!

Share/Save

The dataset, including both articles’ content and social tags, can be downloaded from here.

Share/Save

Abstract

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 the 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.

Share/Save

Abstract

User-generated annotations on social bookmarking sites can provide interesting and promising metadata for web document management tasks like web page classification. These user-generated annotations include diverse types of information, such as tags and comments. Nonetheless, each kind of annotation has a different nature and popularity level. In this work, we analyze and evaluate the usefulness of each of these social annotations to classify web pages over a taxonomy like that proposed by the Open Directory Project. We compare them separately to the content-based classification, and also combine the different types of data to augment performance. Our experiments show encouraging results with the use of social annotations for this purpose, and we found that combining these metadata with web page content improves even more the classifier’s performance.

Share/Save

1. What is Tagging?: As an introduction, the book offers an interesting overview on tagging, letting you discover what it is and its advantages.
2. The Value of Tagging: Why do people tag? Why does a website/intranet need a tagging system?
3. Tagging System Architecture: You will learn that a tagging system involves users, resources and tags in it. Moreover, the relations between them and their features are also presented.
4. Tags, Metadata, and Classification Systems Using tags as metadata, and its differences with a classical taxonomic system.
5. Navigation and Visualization: Advantages of a tagging system for navigation and visualization of a website’s content, showing some new stuff like tag clouds. In this chapter, geotagging is also presented.
6. Interfaces: Some tips on implementing a user-friendly interface for a tagging system. How to ease users to tag a resource, recommending or without recommending tags, how to separate tags (spaces, commas, etc.), and much more.
7. Technical Design: Some technical tips, such as designing the database for a tagging system, and using the open-source tagging plug-in FreeTag to ease this work.
8. Appendix A – Case Study: Social Bookmarking: A brief history and some other ideas on social bookmarking sites.
9. Appendix B – Case Study: Media Sharing: Tagging for rich media, such as images and videos.
10. Appendix C – Case Study: Personal Information Management: How to manage personal information with tags.

Strongly recommended!

Share/Save

Abstract

Las anotaciones generadas por usuarios en sistemas de marcadores sociales pueden proveer metadatos interesantes y muy útiles para la clasificación de páginas web. Estas anotaciones incluyen diversos tipos de información, como etiquetas y comentarios. No obstante, cada tipo de anotación tiene una naturaleza y un nivel de popularidad diferente. En este trabajo, analizamos y evaluamos la utilidad de cada una de estas anotaciones sociales para clasificar páginas web sobre una taxonomía como la del Open Directory Project. Las comparamos por separado a la clasificación basada en contenido, y también las combinamos. Nuestros experimentos muestran resultados prometedores con la utilización de anotaciones sociales para este propósito. Y además indican que su combinación con el contenido textual mejora el rendimiento de la clasificación.

Share/Save