Data for AI-PoR Detection
Description
This dataset was used in the research article "Zero Trust and AI-Driven Dynamic Camouflage for Enhancing Data Security in Cloud-Based Blockchain Centers." It contains the final processed version of the simulated data, prepared specifically for training and evaluating the AI model used in the study. The raw simulated data were preprocessed and are not included in this dataset, as they were not directly used for AI analysis.
Files
Steps to reproduce
The dataset was generated by simulating query activity from 120,000 patients and 250 institutions in Taiwan. The data were preprocessed using Python to apply constraints on query type, timing, and user behavior patterns. The final dataset includes 100,000 legitimate queries and 7,999 malicious queries, with four computed variables based on predefined rules. These variables were used as input features to train and validate an AI model. To reproduce the dataset, one would need to simulate similar entity relationships and apply the same logic and constraints as described in Section 4.3 of the article. Custom scripts used for data preprocessing are available upon request.