Suggestion of method to use

hi all,
I want to do analysis of text data analysis to predict and classify the text into one of the classes.
Which algorithm should be the best to use .
Each text is about 30 lines and contains names ,dates, certain issue related comments.
Corpus size varies from 500-300 documents.
Columns - 10-15

Please suggest the best methods to achieve maximum accuracy.
Thank you

The standard method appears to be LDA; see