Cyberbullying response system on SNS

Abstract

The system is designed to search for cyberbullying-related data and to respond by phase to the cyberbullying on the basis of the data authenticity. In particular, each analyzed data is used for D/B for follow-up cyberbullying data authenticity identification and is stored in D/B for follow-up cyberbullying response, therefore preliminary cyberbullying crime signs will rapidly detect and ultimately, its consequent crime damage will be minimized. In terms of data collection and analysis, the proposed model collects cyberbullying data widely present in various forms, while the previous only collects limited data on Twitter. To complement another limitation of the previous, not being able to analyze video, image and freeform letters, the proposal utilizes SNA to analyze bullying data in freeform letter, image and abstract vocabulary. Furthermore, the existing model does not have DB for follow-up utilities, the proposed model applied a DB for follow-ups for more effective operations of cyberbullying data authenticity identification and response module. Cyberbullying crime response method has also been enhanced in its effectiveness, by enabling phased countermeasures, and storing and analyzing the processed result from response module, to respond to future cyberbullying. Consequently, the proposed model is designed to minimize probable victimized damages by improving effectiveness of rapid and accurate phased measures - via data collection and analysis module, expert system, knowledge-based database, D/B1 for follow-up cyberbullying data authenticity identification, crime response module and D/B2 for follow-up cyberbullying response. The proposed Cyberbullying Response System on SNS enabled an enlarged volume and variety of SNS data collection, compared to the previous model, utilized SNA for analysis, enhanced preliminary prediction accuracy to future cyberbullying and allowed more rapid and adequate phased crime response against it - thus, crime damage from cyberbullying is expected to be minimized.