This is the data set used for the experiments.

Filenames correspond to the filename used on project gutenbarg (www.gutenberg.org), with the exception of the federalist papers in which case they correspond to the order in which the papers where published, on Gutenberg these are available in one combined file.
The original unprocessed text files can be obtained from project Gutenberg, and should be manually stripped of Headers and footers before classification.
The ppLTH is the preprocessed files fed to the LTH parser, and the LTH files are reduced FrameNet info in XML format.(to decrease filesize only frame names were retained, frame elements and other information not used for this project were filtered out.) 