A Weighted GraphSAGE-Based Context-Aware Approach for Big Data Access Control.

Area of Science:

• Computer Science
• Data Science
• Artificial Intelligence

Background:

• Existing context-aware access control (CAAC) methods lack automatic context awareness and the ability to model and reason about context relationships.
• Dynamic access control for big data relies heavily on context information, which current methods fail to fully leverage automatically.

Purpose of the Study:

• To propose a weighted GraphSAGE-based approach for big data access control that enables automatic context awareness and reasoning.
• To address the limitations of existing CAAC methods in automatically modeling and understanding context relationships.

Main Methods:

• Graph modeling of access record data to transform context-awareness into a graph neural network (GNN) node learning problem.
• Development of a GNN model, WGraphSAGE, incorporating weighted neighbor sampling and aggregation algorithms.
• Automatic generation of CAAC rules through GNN node learning and relationship strength modeling.

Main Results:

• The proposed WGraphSAGE model demonstrates superior performance in context awareness and relationship reasoning compared to similar GNN models.
• Experimental results show significant advantages in dynamic access control decisions over existing CAAC models.
• The method effectively achieves automatic modeling and reasoning of node relationships and their strengths.

Conclusions:

• The weighted GraphSAGE approach offers a robust solution for automatic context awareness in big data access control.
• This method significantly improves the accuracy and efficiency of dynamic access control decisions.
• The proposed technique provides a powerful tool for understanding and utilizing complex context relationships in big data environments.