With the emergence of Web 2.0 and the development of social media platforms, more and more users buy products from online pages. After the rise of social media such as Facebook platform, e-commerce business is increasing day by day. Social media is a place where daily number of users increase and lots of users spend lots of time on social media. So, it is a huge scope of market for e-commerce business and very important to perceive the behavior of this type of consumer for being successful in online business. The dataset used in this work collected by a survey which contains total 385 data. This paper analyzed which reason influenced consumers to take the decision to buy or not buy anything from social media. After the survey, found that "Special discounts or offers" is the most common reason why consumers bought anything from social media and "Lack of trust" is the main reason not to buy anything from social media. In this paper, performance of different machine learning methods compared for classification of consumer behavior. The results show that the random forest technique provides the maximum accuracy

Keywords : Consumer behavior, E-commerce, Machine Learning, Social Media, Purchase Intention, Data Mining.