.net Graph Mining

Briefly, I need some advice on a solution that I’m trying to implement. I'm implementing a social network, using .net technologies and I need t开发者_如何学JAVAo infer knowledge from the data collected and persisted at the database (A kind of Graph/Link Mining). My questions are:

1. This task should be performed by a business intelligence tool?
2. Do you know any .net framework to process this type of information (graphs)?
3. How should I persist the result data? Using a separate database?

I'm not a BI expert but yes this does sound like a BI tool could be useful.

I don't know of any .Net library, but CodePlex is usually a good place to look if you're after something open source.

Using a separate database is definitely a good idea if you want to do lots of reporting and the Transactional part of the application is usually busy. FYI transactional databases are usually known as OLTP where-as databases intended for reporting are usually classified as OLAP; the Wikipedia links give a fairly good introduction if the terms are new to you.

if you need automatic inference support you might take a look at Semantic Web solutions http://www.rdfabout.net/intro/? like http://www.dotnetrdf.org/ which provide access to a lot of RDF stores (RDF databases).

Pure graph databases or graph2SQL mappers seldom provide an out-of-the box inference support.

Cheers, Achim