Pedro has a tendency to blog about my favorite subjects, and he knows a lot more about them than I can lay claim to these days, so it shouldn’t come as a surprise that his post on mining Connotea tags caught my attention. Read away to see how a tag-based search and associations using that yield some rather cool results.
This is a simple example, and obviously Connotea and CiteULike are still far from being globally accepted repositories for scientific knowledge (hopefully that will change in the near future), but the potential of such sites is quite clear. In my own research, be it related to my work or other searches, I use del.icio.us quite heavily, and it is probably still the most useful of all “web 2.0″ sites. With biology, the ability to mine text and build associations and relationships only makes things more interesting. Earlier today I was reading an article in Genetic Engineering News about AKS2, a text mining/knowledge discovery application. How successful would an application like that be if it was applied in the kind of “social” context that Pedro used in his experiment? I wonder if anyone who makes a living in this field has given the matter some thought.
Further reading
An open scientific future
Biology, search and Udell
Biological content: Access and monetization
The future of science is open, part 3 (read the first two parts as well)
Folk wisdom for web sites
(Picture Source: ycc2106)
Technorati Tags: Search, Connotea, Data Mining, Bioinformatics
4 Comments
I should have really stressed more that I am not really familiar with text mining
.
It is interesting to think of sub spaces in the social web. If had tried to do this in delicious I would probably get a very different result since Connotea is much more focused on science and annotation of science papers. In a way the ideal situation would be to have a global tag space but more information on the taggers. One could decide to use or not the tags according the individual’s knowledge base.
I should have really stressed more that I am not really familiar with text mining
.
It is interesting to think of sub spaces in the social web. If had tried to do this in delicious I would probably get a very different result since Connotea is much more focused on science and annotation of science papers. In a way the ideal situation would be to have a global tag space but more information on the taggers. One could decide to use or not the tags according the individual's knowledge base.
It’s the idea that counts.
Sub-spaces are fascinating. Since information is distributed and not equal everywhere, how do we judge it? Jon Udell and others talk about this “level of trust”, which doesn’t only apply to people. Maybe each knowledge base could be given some sort of trust score, or the system learns the trust score somehow.
It's the idea that counts.
Sub-spaces are fascinating. Since information is distributed and not equal everywhere, how do we judge it? Jon Udell and others talk about this “level of trust”, which doesn't only apply to people. Maybe each knowledge base could be given some sort of trust score, or the system learns the trust score somehow.