Support Joomla!

Wed

30

Apr

2008

Categorizing and linking Joomla article in a meaningful way
User Rating: / 2
PoorBest 
Written by Danh Le Phuoc   

Have you ever think that "is it possible to categorize and link your Joomla! articles with entities (people, places, organizations, etc.), facts (person ‘x’ works for company ‘y’), and events (person ‘z’ was appointed chairman of company ‘y’ on date ‘x’)?".Yes, theorically, it's possible,but, to have such functionalities, we have to spend a lot of efforts on natural language processing, machine learning and other stuffs. So,practically,not feasible for a small and medium CMS. However, it's still insteresting and worth trying,isn't it? .So,let's try a webservice like Calais[1] which provides such funtionalities through websevice calls.

In Calais Overview site they said that "The Calais web service automatically attaches rich semantic metadata to the content you submit" and  such "metadata gives you the ability to build maps (or graphs or networks) linking documents to people to companies to places to products to events to geographies to … whatever. You can use those maps to improve site navigation, provide contextual syndication, tag and organize your content, create structured folksonomies, filter and de-duplicate news feeds or analyze content to see if it contains what you care about. And, you can share those maps with anyone else in the content ecosystem". This sounds very attractive,right? . But don't be too optimistic, it is not perfect yet. Let play arround with their viewer at [2], to find out yourself which kind of extracted data they can give you. Or you can check out [3] to  see very useful comments of their usecase. And actually, Wordpress and Drupal already integrated Calais to enable auto-tagging, see the Gallery section of Calais website to interesting examples. Let's come up with an extension or plugin for Joomla, it won't take much effort,but definitely could be a lot of fun with it!!!!

[1] Calais.

[2] Calais Viewer.

[3] Powerhouse Museum 

 

 

 

4 Comments

  1. Wilco Jansen makes this comment

    Hi Danh, I reall like your project, and also like to see you start blogging here. Love to see your project start, will certainly keep a close eye to it. Wilco

    Wed 30 Apr 2008 19:38:42 EDT

  2. Aini makes this comment

    Great Post ! You did it ...

    Wed 30 Apr 2008 22:17:50 EDT

  3. Amy Stephen makes this comment

    OK. That *is* cool. So, how about this idea? 1. Integrate Dan Grossman's Open Calais Tags PHP class. [1] 2. Then, create a plugin that fires on "onPrepareContent" shipping the article to Open Calais, returning an array of people, companies and terms identified in the article. 3. Feed that list into a Technorati Searchlet.[2] 4. Provide list results with links to "related blog postings" on the people, businesses and things found in your article. http://www.dangrossman.info/open-calais-tags/ http://extensions.joomla.org/component/option,com_mtree/task,viewlink/link_id,3887/Itemid,35/

    Wed 30 Apr 2008 23:21:04 EDT

  4. Danh Le Phuoc makes this comment

    Cool,that's good idea, Amy.I'll find the way to play arround with it,anyone is interested in it,give me a vote or let me know.BTW,I'll have a post about "related postings" with Semantic Web search engine and Joomal!RDF data exposing. In this post you'll see the way that we can have "Interconnected messages and topics" on forums,wikis,mailing list,IRC archives,CMS,etc through some simple APIs. rdfs:seeAlso : http://www.sindice.com/dev?section=siocwidget :)

    Thu 01 May 2008 01:58:26 EDT

Add a Comment