Tagging With Power{Comments Off on Tagging With Power}


by James C
Summary:

Folksonomies are used by numerous Web 2.0 services and have evolved into a vital user-driven tactic to index and retrieve information. The indexing of information is a strength of folksonomies. However, it is its retrieval methodology that is its weakness. Power tags represent a new search option that improves the type of retrieval findings. There are two kinds of tag distributions when dealing with the application of power tags. The first type of tag distribution is the power law distribution. The power law distribution has a mathematical expression whose curve simply contains a limited number of tags at the top of the distribution. It also has several tags with lower ranks on a long tail on the right side of the curve. The second type of tag distribution is known as inverse logistic distribution. Inverse logistic distribution contains many applicable tags at the start of the curve and is also called a long tail. Derived power tags are used to limit the number of searchable tags. The purpose of limiting searchable tags is to limit the number of search results and to improve accuracy. Further research will need to focus on how to take the theoretical ideology and turn it into actual solutions that assesses retrieval outcomes founded on power tags.

Reflections:

I really enjoyed this article because it gives  insight into what may be happening on the horizon in terms of power tags. I totally agree that while the indexing process runs quite smooth and is very efficient, the retrieval process can definitely be improved. The fact that derived power tags can simultaneously limit the number of search results yet be efficient is something I definitely look forward to.

Moreover, I liked the fact that the article dealt with a theoretical topic that can actually be converted into an actual process. Many times in reading theoretical and technical articles I am unable to really see how a theory can be turned into reality. However, this article takes the reader step by step as to how this is possible.

Citations:

Peters, I., & Stock, W. G. (2010). “Power tags” in information retrieval. Library Hi Tech, 28(1), 81-81-93. doi:10.1108/07378831011026706