The proliferation of artificial accounts, usually referred to as “bots”, on social mediaplatforms poses major challenges to the authenticity and reliability of online discourse.Consequently, detecting and mitigating the influence of these bots has become a criticaltask to preserve the integrity of social interactions online. Motivated by the needfor effective bot identification, this thesis focuses on the TwiBot-22 dataset, a recentbenchmark comprising an extensive array of human and bot user profiles from Twitter.The objective is to develop an effective system for bot identification. Two key goals of theresearch are to evaluate the effectiveness of machine learning models in spotting socialbots and to examine the performance of the implemented models with results reportedin the existing literature. The results help to clarify the strengths and shortcomingsof current bot identification techniques and offer suggestions for future developmentsin this domain, thereby improving the knowledge of bot activity and bot detection approaches.