Improvement of the robustness and the integration of subtyping analyzes by computational statistics and software integration

Scientific panel: Life sciences
Total duration in months: 15
Call identifier: FP7-PEOPLE-IEF-2009
Keywords: data mining; computational statistics; statistical inference; R packag- ing; mixture modeling; cluster analysis; visualization; subtyping; clinical research; complex pathologies

Mayweb is a local search engine started in the late 90's by Philippe Minzière. At that time, it aimed to reference the web ressources in the local french region of Mayenne.

Maintaining in my spare time the Mayweb local search engine, I face the burden of spam weblink submission. The problem is such that about 99.9% of the submitted links are spams. To address that issue, as the web form was previously unprotected, I first added a request to read the code of an image.

But this did not solve the whole problem as many submitters were still capable to send unrelated links to our directory.

Patrick Richard is a teacher of technology in a secondary school in the west of France. Patrick is always looking for new ways of translating recent developments of IT into his teaching.

As part of this perpetual research, he funded in 2000 (with others) the non-profit making organization pagestec.org, whose aim is to favour and federate the collaborative use of Internet as a development tool in technology teaching in secondary school. Nowadays, the association federates several 100 of teachers around the association, its website and mailing list.

So, knowing each other for quite a few years -as he actually taught me a few years back!-, he invited me to make a proof of concept (PoC) search engine based on a collection of websites related to technology teaching in France.