========================================================== Datasets as used in CINLP special issue paper submission ========================================================== The zip archive contains the following files: - smiley_trainingSet.csv - Contains training data that was labeled automatically using smiley as an indication of the sentiment. It is used for training RBEM. - newTruth_trainingSet.csv – Contains training data that used to expand the RBEM models in addition to the smiley data. The file is a CSV with the following fields: CLEANED_MESSAGE,POS_TAGS,LANGUAGE,SUBJECTIVITY,POLARITY, Where SUBJECTIVITY is either 1 (subjective) or 0 (objective) and POLARITY is one of {pos,neg,obj} - newTruth_testSet.csv – Contains test data upon which the polarity classifiers were tested. The file is structured in the same way the trainin set file is.