Author : Vaishali Govind Bharane and Prof. Bere Sachin S. Assistant Professor

In the present generation number of clients can communicate with each other through social networking sites such as Facebook, Twitter, WhatsApp, etc. The social networking sites are used in the world a huge number of clients can communicate with each other. Online Social Networks (OSNs) have become increasingly popular. People’s lives have become more associated with these sites. People are used to Online Social networks to keep in touch with each other and communication between social networks for share news, organize events and advertisement of own e-business. The increasing growth of OSN and the more amount of personal data of its subscribers have attracted attackers and imposters to steal personal details, share fake news and spread malicious activities. On the other side, researchers have started to research efficient techniques to detect abnormal activities and fake accounts relying account on different features and classification algorithms. However, some of the accounts exploited features have a negative contribution in the final results or it has no impact it has using standalone classification algorithms does not achieve satisfactory results. In this paper, we present a machine learning technique to identify fake accounts on twitter. We have a preprocessed dataset of numerical highlights. The Support Vector Machine algorithm is proposed to provide efficient detection of fake accounts of twitter it has used feature selection and dimensionality reduction techniques. The machine-learning algorithm was used to decide accounts to identify accounts that are fake or real. SVM algorithm is used to identify account is fake or real. SVM has used a smaller number of features hence it is being able to correctly classify about 98% of the accounts of our provided training dataset.

Keywords: Machine learning; online social media; Twitter