Enhanced Detection of Social Network Spam Using Improved Extreme Learning Machine

This research employs an Improved Extreme Learning Machine (IELM) algorithm to address the issue of spam identification in social networks. By fine-tuning activation functions and improving weight initialization, the IELM improves on the conventional Extreme Learning Machine (ELM), enhancing its performance and capacity for generalization in high-dimensional data contexts, such as social networks. By categorizing characteristics like posting frequency, link usage, and message content, the system can identify spam accounts. The IELM model minimizes false positives and produces excellent accuracy when trained on a variety of social media post datasets. This technique protects user experience and platform integrity by assisting social network platforms with real-time spam identification.