We developed a model for sentiment classification. Validation was performed on three datasets. Bagging, Naive Bayes and Support Vector Machines were used. We showed that multiple classifier systems increase the performance of sentiment classification.