Now a days Most of the people using social networking sites regularly. Users may create accounts and engage with others at any time and from any location, making social networking services increasingly popular. There are several benefits of using social media, but disclosing personal information onlinecarries hazards as well. There are also risks associated with sharing personal information online. We need to categorize the user identities of online communities in order to analyze who is promoting dangers in these platforms. The categorization allows us to determine which social media profiles are authentic and which are fraudulent. Typically, Different categorization approaches have been used to recognize fake identities on social networking platforms. However, we need to enhance the reliability of fraudulent profile identification in online communities. In order to boost the trustworthiness of the false profile identification, in our paper we are going with Machine Learning (ML) technologies and Natural Language Processing (NLP) methods for this, we opted KNN, Naive Bayes and Support Vector Machine (SVM) algorithms in our work