[TYPO3-Solr] Explanation of scoring wanted

[Jigal van Hemert]

Hi,

On 09/03/2015 19:50, Alexander Stehlik wrote:
> I can not be sure yet, because the nutch DB is still building up and
> maybe the results will change again over time, but at first sight, this
> is what worked for me:
>
> 1. Remove the scoring-opic string from the plugin.includes config in
> nutch-site.xml

I asked the Nutch community about this and the OPIC reference seems to 
ring some bells. The concept is described in an article [1] which also 
explains a bit of idea behind it. OPIC (and especially the Adaptive 
variant) adds a boost to the documents based on their relationship with 
other documents in the index and the weight of those related documents. 
In short, it can be compared a bit with the page rank mechanism of 
Google; if a lot of documents are related to a document it must be more 
important than just any document and if important documents are related 
to a document then that document must also be a bit more important than 
the rest.

It seems that the documents from TYPO3 do not receive this treatment. 
Therefore it would be wise to remove scoring-opic from the plugin list.
In recent versions this has also been removed (for a different reason).

An issue for this is created [2].

[1] http://www2003.org/cdrom/papers/refereed/p007/p7-abiteboul.html
[2] https://forge.typo3.org/issues/65658

-- 
Jigal van Hemert
TYPO3 CMS Active Contributor

TYPO3 .... inspiring people to share!
Get involved: typo3.org