text clustering algorithm (in Java)

I need a text clustering algorithm implemented in Java. You need to suggest an algorithm to be used and the reason why. Your bid price on this specific project is for the creation of a prototype only with the purpose of demonstrating the working of the algorithm. The eventual application is to be written in a way that i can use it in multiple situations and websites. So ideally i want to be able to submit a collection of text to process and the application returns clusters with references to the original text submitted. Both submitting and returning results would ideally be done using a MySQL database. A variation may be to also cluster texts into groups of pre-defined keywords. We will discuss the implementation of the eventual application options at length, this project is about finding a startingpoint and prototype. Examples of implementations are: Google news: [login to view URL];cf=all&ned=us&topic=s&ict=ln where news items are clustered by subject. Clusty search: [login to view URL] where results of a search engine are clustered The first website where i want to implement this is: [login to view URL] , the news items from RSS feeds are stored in a MySQL database and a search can be done based on Title + Description + Content . You do NOT need to do any work on the website itself, only the programming of the clustering algorithm. So for the prototype i want to submit all the Title lines from the past 3 days and receive the clustering on that group. That may be a fairly large group but it can be split based on publisher or sport should that generate more efficient results.