Hi all,
I am a 2nd yr undergrad student studying CS at Bengal Engineering and Science University, India. This is the first time I am participating in Google Summer of Code. And I am very interested to contribute to machine learning or NLP related projects.
I went through the Python projects of gsoc 2013 and I found the anti-spam/anti-abuse filter of Mailman very interesting. I had some own ideas regarding the project. I want to integrate/use external NLP toolkits like Lingpipe and LIBSVM. Classification of the text will be done on an N-gram Language Model and a Support Vector Machine (I would prefer LIBSVM), with the "localized knowledge" (which would be used as the training set for the classifier) to help in identifying the abuse/spam. Adding a feedback procedure to the classifier might also be helpful as it will help us to improve classification and the classifier can update itself with previously classified spams/abuse and hence the program can filter latest spams/abuses without any update, on programmer's part.
About me: programming club.
- About 5 years of programming experience mainly in Java,python, C and C++ and a bit of PHP and LISP. I am a part of our college Software and
- I qualified the Zonal Informatics olympiad 2010-11. I regularly participate in online coding competitions and hackathons.
- Developed a email client in python.
- I have some prior experience in machine learning and natural language processing. I participated in Twitminer 2013 (IISC Bangalore) where our team (BEing) came 6th. In the last few months, I have been working on the sentiment analysis of tweets, based on Language Models of Lingpipe Toolkit of Java.and LIBSVM to analyze the efficiency of classification.
Looking forward to hear back from all of you!
Pratik Sarkar