Optimized adaptive intrusion detection framework for big data in social media application

Social media has become a significant aspect of individuals' everyday lives as it enables communication and information exchange. However, these channels are also being used to spread misinformation and harm others. A novel approach called the Coati Attention Transformer Prediction (CATP) framework was implemented, where social network intrusion detection data was initially considered and trained in the system. Preprocessing was conducted to eliminate noise variations from the trained database, and present features in the database were estimated by the coati optimal behavior. Then, the attack was predicted, and classification was conducted based on different classes. The suggested model exhibits an outstanding success rate, achieving an F1-score of 99.98%, a recall of 99.98%, a precision of 99.98%, an error rate as minimal as 0.02%, and an accuracy of 99.98%. The proposed model shows the best success rate with accuracy.