How should I use mlxtend for sentiment analysis?

Question

I have created classifiers like naive bayes, logistic regression and SGDC, how should I use this ensemble voting method, to create a final score, I know there are both hard and soft voting. My featuresets are words basically.

eclf =  EnsembleVoteClassifier(clfs = [clf1, clf2, clf3], voting = 'hard')
eclf = eclf .fit()

How should I proceed?