ABSTRACT
Artificial intelligence (AI) is a widely used technology to solve the issues of contemporary cyber threats on social media platforms. Natural language processing (NLP) is a subset of artificial intelligence that significantly helps prevent cybercrime and restrict access to malicious social media posts. Social media post analysis (SMPA) helps analyze the interactions among different people/groups on social media platforms, such as Twitter, Facebook, Instagram, and Reddit. SMPA aids in investigating revolutionary plans and keeping an eye on terrorist activities through social media posts, chats, and offensive posts exchanged among the rioting community against societal norms. Social media networking has become a new target for cybercrime. Cybercriminals are trying to exploit people’s innate trust in their social media networks by spreading malicious software, links, and fake transaction links and sending spam messages. This chapter provides significant ideas for preventing cyberattacks on social media networks using a better identification approach in named entity recognition. Named entity recognition (NER) identifies and extracts meaningful information from unstructured data, such as Facebook and Twitter comments using different techniques such as rule-based NER, machine learning-based NER, and hybrid NER. We propose a CyNER system to recognize entities like I-Malware, O (others), B-Malware, B-System, I-System, B-Organization, B-Indicator, I-Organization, I-Indicator, B-Vulnerability, and I-Vulnerability from a given social media post containing cybersecurity issues using random forest and different upgraded RNN architectures such as long short-term memory (LSTM) and BiLSTM. The experimental results show that the proposed CyNER system’s accuracy is better than BiLSTM and the random forest algorithm.
