Last updated May 7, 2006.
Short Description:
Should be able to build automated context of nodes. This could be: related articles, language detection (e.g).
Motivation:
To build social and intelligent software is one of the main goals.
By classifing nodes (content) into a content-context, you would be able to realize automated language or related conent detection and for example content assignment to categories. This would save a lot of work and has a objective point of view.
This idea is a subpart of some definitiones defined in Web2.
Success criteria:
The module should be recognize the relevant words (context) of an article. This should be stored in a table.
Nonrelevant word should be detected by mechanism, so that the context.table is as small as possible.
There should be some "score" definitions, which decide the best context-score by:
- relevance
- age
- length of article
Module should be able to generate:
- Related content
- Meta Keywords
- Category assignment
- Language detection
Roadmap:
It's related to several modules like spam. nodewords, whatsrelated, search, similar entries and so on.
First step should be to build the context tables
Second step to bring content in relation
Close to this is the project Content recommendation engine