Published Paper Details:

In today's world, Online Social Media is king in a number of forms the number of people who use the service is growing every day. The use of social media is skyrocketing. The primary benefit is that we can easily communicate with people via online social media and communicate with them in a more effective manner. This opened up a new avenue of a possible attack, such as a forged identity, false information and so on. According to a recent study, the number of accounts in the number of people who use social media is much higher than the number of people who use it. These fake accounts are difficult to detect for online social media providers. Since social media is flooded with false information, ads, and other types of content, it is essential to recognize these fake accounts. From an online social media dataset, we offer a method for detecting fraudulent accounts. We employed boosting methods to improve the accuracy of the standard technique, rather than employing typical machine learning classifiers. By boosting weak learners, this method has resulted in a large improvement in accuracy. In this paper we will use accuracy comparison of Xgboost Classifier, and Gradient boosting Classifier. Xgboost performed brilliantly when compared with the previous work.