Most people use social networking sites in their daily lives these days. On social networking sites, several people create their profiles every day and engage with others regardless of their time and location. The social media platforms offer their users benefits, but they also raise concerns about their information security. Categorizing user profiles in order to determine who is endorsing threats on social media is necessary. The categorization helps us discern between authentic and fraudulent social network profiles. Traditionally, there have been a variety of categorization techniques used to identify fraudulent social network profiles. However, we must raise the accuracy rate at which false profiles on social networks are identified. In order to increase the false profile identification accuracy rate, we provide machine learning and natural language processing (NLP) methodologies in this work. Support Vector Machine (SVM) and Naïve Bayes approach are available.