I wrote two perceptron classifiers (vanilla and averaged) to identify hotel reviews as either true or fake, and either positive or negative using word tokens as features.
The classifier will simply ignore unknown tokens in the test data.
Mapped punctuations to space and lowercased all the letters.
Ignored stop words.
Compare the output generated by my model ('percepoutput.txt') with 'tagged_devdata.txt' from the data folder.