A Supervision Method for Online Violence Based on Blockchain and Deep Learning

With the burgeoning development of the internet, online violence has metastasized into a formidable concern impeding cybersecurity and societal stability. This research proposes an innovative supervision methodology for online violence based on blockchain and deep learning technologies. The proposed technique harnesses the security, credibility, and traceability attributes of blockchain to ensure data integrity and authenticity alongside the efficiency, accuracy, and robustness of deep learning models for expedient text classification. Extensive experimentation on a 50,000-sample dataset validates that the introduced BCLD approach attains 0.95 accuracy and F1 score, outperforming baseline models by 4–5%. Additionally, balanced handling of positive and negative samples is demonstrated. The tamperproof nature of blockchain thwarts fraudulent data alteration. This pioneering methodology signifies a paradigm shift for constructing a wholesome online ecosystem.