Network Location-Aware Service Recommendation with Random Walk in Cyber-Physical Systems

Area of Science:

• Computer Science
• Engineering
• Information Systems

Background:

• Cyber-physical systems (CPS) increasingly rely on service-based functions.
• Traditional service recommendation methods, like collaborative filtering (CF), face challenges with data sparsity and context.
• Accurate service recommendation is crucial for efficient CPS operation.

Purpose of the Study:

• To propose a novel service recommendation method for CPS.
• To address the limitations of traditional CF methods in sparse data environments.
• To leverage contextual information, specifically network location, for improved recommendation accuracy.

Main Methods:

• Developed a novel service recommendation approach for CPS.
• Utilized network location as a key contextual factor.
• Implemented three prediction models based on random walking techniques.
• Mined potential similarity relations among users and services.

Main Results:

• The proposed method demonstrates effectiveness in service recommendation for CPS.
• Incorporating network location significantly improves Quality of Service (QoS) prediction accuracy.
• The approach mitigates issues related to data sparsity inherent in traditional CF methods.
• Experiments on real-world datasets validate the proposed models' performance.

Conclusions:

• Network location is a valuable contextual factor for enhancing CPS service recommendations.
• The proposed random walking-based models offer a promising solution for accurate QoS prediction in CPS.
• This research contributes to more efficient and reliable service discovery in complex cyber-physical systems.