Sentiment analysis is the process of determining opinion expressed in a text, or an estimation of emotion related to the certain topic if it is negative, positive or neutral. The massive growth of social media, Twitter has played an important role since it allows people to express their feelings about a subject. Classification algorithms are necessary in the process of sentiment analysis. In this paper, we build a model to acquire people’s opinion on any concerning subject and evaluate the classification algorithms on the dataset. To accomplish the goal, we use a large set of Tweets which refer to a particular topic and execute the analytics on the Twitter feeds to classify them by using Naïve Bayes, Support Vector Machine, Maximum Entropy and Boosting algorithms. Then, to obtain the result we measure the accuracy among the four algorithms and compare them to identify the best algorithm based on our experiment.