 This archive has the following directory structure:

	sle_convote \
               |-- ./speaker-based
               |-- hiddenvars_OpinionFinder.txt
               |-- README.txt
               |-- sample_script.sh
               |-- ./bin
               |-- feature_convote.txt

 -- ./speaker-based contains the data in a format that SVM^{sle} software requires. It training/validation/test set. The experimental setup used for this experiment is exactly as described http://www.cs.cornell.edu/home/llee/data/convote/README.v1.1.txt

 --  hiddenvars_OpinionFinder.txt contains the initializations  of Latent Explanations with OpinionFinder for the training file.

 --  README.txt -- current README.txt

 -- sample_script.sh  is a sample bash script that shows how to run a few interations of SVM^{sle} on U.S Congressional Floor Debates data. 

 -- ./bin folder  should  contain  executables for SVM^{sle}. Download source code of SVM^{sle} from 
    
                                              http://projects.yisongyue.com/svmsle/ 
                                              
    and compile it, then place executables to the ./bin folder.
  
 -- features_convote.txt  file contains the interpretation of each feature on the corresponding line, for example the meaning of feature 100 is on the line 100.
