Sentiment Analysis using perceptron

Question

I am trying to implement Sentiment analysis using perceptron to get a better accuracy in python. I am lost in the maths that sorounds it and need easy explanation on how to port it to be used for sentiment analysis. There is already a paper published on the same : http://aclweb.org/anthology/P/P11/P11-1015.pdf

Would anyone here be able to explain in detail and clarity ? I have a training datatset and test dataset of 5000 reviews each and am getting an accuracy of 78 percent with bag of words. I have been told perceptron will give me an accuracy of 88% and am curious to implement it.

Answer 1

Perceptron is just a simple binary classifier, that works on fixed size vectors from R^n as input data. So in order to use it you have to encode each of your documents in such a real-valued vector. It could be for example a bag-of-words representation (where each dimension corresponds to one wor, and the value to number of occurences), or any "more complex" representation (one of which is described in the attached paper).

So in order to "port" perceptron to sentiment analysis, you have to figure out some function f, that feeded with document returns real-valued vector, and then train you perceptron on pairs

(f(x),0) for negative reviews

(f(x),1) for positive reviews