An Optimizing Crime Detection in Social Media Platforms Using Multiagent Ontology-Based Approach

Social media platforms have grown exponentially in recent years, creating both opportunities and challenges for crime detection. As criminal activities and suspicious behaviour often manifest through social media content, developing efficient and effective methods for detecting them is essential. The ontology-based agents utilize semantic analysis and natural language processing to identify suspicious patterns and structures in the data. The multi-agent optimization algorithms then utilize this data to optimize the detection process and improve accuracy. This optimization helps to improve the system's efficiency, response time, and resource allocation, the obtained results show that the proposed optimization framework offers improved detection accuracy, faster response times, and better resource allocation compared to traditional approaches. The existing approaches like The Naive Bayes model achieves an accuracy of 78.69%, and the Decision Tree model has the lowest accuracy of 74.78%. Overall, our experiments and comparative analysis Demonstrate the effectiveness of the proposed optimization framework. The results of our analysis show that our optimization framework excels in detection accuracy, response times, and resource allocation compared to traditional approaches. The proposed methods accuracy value is 99.01%, precision value is 98.87%, f1-score value is 98.76, and ROC 98.87%. This reinforces the efficacy of our proposed approach and further validates its potential for use in practical scenarios.