We are looking to add automated analysis of articles to our web-based platform and are currently tossing up between licensing a platform or building a custom solution.
We are looking for the following requirements:
1. Automatic classification of articles on topics
2. Automatic identification of companies
3. Automatic identification of people
4. Automatic identification of products
5. Analysis of prominence of company/person/product mention (Prominent, brief, throw-away)
6. Analysis of sentiment of company/person/product (positive/neutral/negative)
7. Automated summarisation of articles
8. Matching and clustering of articles
9. Automatically creating profiles of authors based on articles published
You should have proven expertise in this area and references. You should indicate which library you would use and relevent licenses/costs (we are willing to pay for a license for something like lingpipes).
Obviously there is much more detail needed and we only expect ballpark figures. Our platform is built on PHP/LAMP.