ekt1 / valence Goto Github PK

Project Statistical Models of the Emotionality of Speech and Written Text (2011-2014) was supported by the National Programme for Estonian Language Technology (2011-2017) of the Estonian Ministry of Education and Research.

This is a program classifying an Estonian text as positive, neutral or negative. It is running on server http://peeter.eki.ee:5000/valence

This code is optimized for server environment but can be run also as a standalone program:

python valencecolor.py textfile.txt

This will produce a file textfile.txt.html where all emotionally relevant word are marked with CSS style. The limitation is that it does not print out the summary information. As a prerequisite the NLTK must be installed on the computer.

Files:

• valence/korpus.csv - corpus for training the NLTK model
• valence/sqnad.csv - dictionary of Estonian words
• valence/bayes.py - NLTK wrapper
• valence/valencecolor.py - main program

If you use the corpora valence/korpus.csv, valence/sqnad.csv in your work, please cite the following paper:

Pajupuu, Hille; Altrov, Rene; Pajupuu, Jaan (2016). Identifying polarity in different text types. Folklore. Electronic Journal of Folklore, 64, 25−42. DOI PDF