Probabilistic latent semantic analysis(PSLA)
First of all, hello everyone on the Java forums,
I am new here and still studying but I am not an inexperienced programmer. I originally come from C# but I really wanted to try out the UIMA framework and mainly for this question: How do I implement Probabilistic latent semantic analysis?(Find it a really interesting subject(see Watson :D) and will be part of a future school project.)
I searched the forums and the only information I found was this:
http://www.java-forums.org/advanced-...ys-string.html
Which does not really help me any further but I found semanticvectors - Package for creating and searching semantic vector indexes by wrapping Apache Lucene. - Google Project Hosting
Problem is that they focus mainly on the "latent semantic analysis" and not on a Probabilistic version(or any evolved version of PSLA).
So my question is does anyone in here has experience with this method of Natural Language Processing and if so, where do I start to program it?
Dont get me wrong, I know it is a BIG learning curve to fully understand a method like this(+machine learning where I am familiar with) and get familiar with using it. Therefore I dont expect a ready answer and the project to take months to finally create a final version. However since I could not find any code samples or explanations this would be a good place to start and share your/my knowledge about it!
Greets, Olaf
