Trained Spam Machine Learning Classifier/Model

I have a list of sentences about 17 million. I need to 开发者_JAVA技巧identify sentence as spam/ham/unsure. Are there trained models present on the internet to which I could just feed in my data as a "test" set and the system would classify my sentence as spam/ham ?

Note: The sentences aren't e-mails.

You can use the Bayesian Spam Filtering, take a look at this article pretty nice to understand the theory http://robotics.stanford.edu/users/sahami/papers-dir/spam.pdf