This paper introduces a novel approach for automatically classifying the sentiment of Twitter messages. The goal is to classify tweets as either positive or negative with respect to a query term. This is useful for consumers who want to research the sentiment of products before purchase, or companies that want to monitor the public sentiment of their brands. The paper presents results of machine learning algorithms for classifying the sentiment of Twitter messages using distant supervision. The training data consists of tweets with emoticons, which are used as noisy labels. This type of training data is abundantly available and can be obtained through automated means. The paper shows that machine learning algorithms (Naive Bayes, Maximum Entropy, and SVM) have accuracy above 80% when trained with emoticon data. The main contribution of this paper is the idea of using tweets with emoticons for distant supervised learning. The paper discusses the characteristics of tweets, including their length, data availability, language model, and domain. It also describes the approach used, which involves different machine learning classifiers and feature extractors. The paper evaluates the performance of these classifiers on a test set of tweets. The results show that the classifiers perform well, with accuracy above 80% for Naive Bayes, MaxEnt, and SVM. The paper also discusses the use of emoticons as noisy labels and the impact of feature reduction on the performance of the classifiers. The paper concludes that using emoticons as noisy labels for training data is an effective way to perform distant supervised learning. Machine learning algorithms can achieve high accuracy for classifying sentiment when using this method. The paper also discusses future work, including the use of semantics, domain-specific tweets, handling neutral tweets, internationalization, and utilizing emoticon data in the test set.This paper introduces a novel approach for automatically classifying the sentiment of Twitter messages. The goal is to classify tweets as either positive or negative with respect to a query term. This is useful for consumers who want to research the sentiment of products before purchase, or companies that want to monitor the public sentiment of their brands. The paper presents results of machine learning algorithms for classifying the sentiment of Twitter messages using distant supervision. The training data consists of tweets with emoticons, which are used as noisy labels. This type of training data is abundantly available and can be obtained through automated means. The paper shows that machine learning algorithms (Naive Bayes, Maximum Entropy, and SVM) have accuracy above 80% when trained with emoticon data. The main contribution of this paper is the idea of using tweets with emoticons for distant supervised learning. The paper discusses the characteristics of tweets, including their length, data availability, language model, and domain. It also describes the approach used, which involves different machine learning classifiers and feature extractors. The paper evaluates the performance of these classifiers on a test set of tweets. The results show that the classifiers perform well, with accuracy above 80% for Naive Bayes, MaxEnt, and SVM. The paper also discusses the use of emoticons as noisy labels and the impact of feature reduction on the performance of the classifiers. The paper concludes that using emoticons as noisy labels for training data is an effective way to perform distant supervised learning. Machine learning algorithms can achieve high accuracy for classifying sentiment when using this method. The paper also discusses future work, including the use of semantics, domain-specific tweets, handling neutral tweets, internationalization, and utilizing emoticon data in the test set.