Recent Papers

The rising occurrence of fraudulent actions on internet-based social platforms poses a substantial problem for ensuring secure and reliable online environments. This review paper examines the progress and application of artificial intelligence (AI) frameworks that employ machine learning (ML) approaches to detect fraudulent profiles on social media. We conduct a thorough analysis of current literature to emphasise the main approaches, algorithms, and measures of effectiveness employed in cutting-edge artificial intelligence systems designed for detecting fraudulent activities.The paper explores different machine learning methodologies, such as supervised and unsupervised learning, that are used to identify abnormalities and forecast fraudulent activities. We analyse the importance of feature extraction techniques, specifically in the context of user behaviour analysis, interaction patterns, and content characteristics. These characteristics include activity frequency, network connectivity, language features, and multimedia material. The research also investigates the use of hybrid models, which merge decision trees, support vector machines (SVM), and neural networks, in order to improve detection accuracy and resilience.Our research shows that the AI frameworks with the highest efficacy attain accuracy rates of over 95% in detecting bogus profiles. Moreover, these frameworks showcase their ability to adjust to changing fraudulent strategies by employing ongoing learning mechanisms. The paper also discusses the difficulties of detecting fraud in real-time and emphasises the importance of having varied and extensive datasets to enhance the ability of models to perform well across various social platforms and types of fraud.To summarise, this analysis highlights the capacity of AI-powered solutions to reduce fraud risks on social platforms. This information is essential for academics and platform administrators who want to create more advanced and flexible fraud detection systems. Potential areas for future research involve incorporating real-time detection capabilities and broadening datasets to encompass a wider range of social platforms and types of fraud.

Copyright © 2024 JAY KUMAR1. This is an open access article distributed under the Creative Commons Attribution License.