Millions of individuals use Twitter, one of the most prominent social media sites, to exchange information, broadcast tweets, and follow other users. Twitter being an open application programming interface is vulnerable to attacks from fake accounts. Fake accounts are primarily used for advertising and marketing, defamation of an individual, consumer data acquisition, increasing fake blog or website traffic, sharing wrong information, online fraud, and control. Fake accounts are disruptive to both users and service providers, thus it's critical to recognize and filter out such information on social media. This paper presents a technique for detecting fake Twitter accounts using the feature set mapped to Twitter’s rules and policies to flag suspicious accounts. To choose the apt subset of features from the original feature space, feature selection techniques such as information gain and correlation were used. Users have been classified using a Logistic Regression classifier with the detection accuracy of 93.5 percent after being trained on the data of about 1,00,000 Twitter users. Furthermore, an algorithm for classifying suspicious users as Fake profiles has been presented. The results of the experiments show that the suggested model outperforms competitive state-of-the-art research in the majority of circumstances.