Research on the integrated system of case-based reasoning and Bayesian network

Area of Science:

• Engineering applications
• Artificial Intelligence
• Data Science

Background:

• The big data era necessitates efficient intelligent systems for engineering.
• Integrating Case-Based Reasoning (CBR) and Bayesian Network (BN) shows promise but faces challenges with large datasets.
• Existing methods struggle with parameter management and computational efficiency in big data environments.

Purpose of the Study:

• To improve the efficiency and accuracy of integrated CBR and BN systems for big data.
• To develop a parallel data processing strategy suitable for Hadoop platforms.
• To enhance the probability learning capabilities of intelligent engineering systems.

Main Methods:

• Proposed the Within-Cross (WC) algorithm for efficient big data assignment to Hadoop slave nodes, enabling parallel processing.
• Introduced the Weighted Super Parameters of Dirichlet Distribution (WSPDD) algorithm for advanced probability learning.
• Utilized a case study from an application domain to validate the proposed methods.

Main Results:

• The WC algorithm significantly reduces processing time by optimizing resource utilization on the Hadoop platform.
• The WSPDD algorithm improves probability learning accuracy by weighting super parameters of the Dirichlet Distribution.
• The integrated system demonstrates enhanced efficiency and accuracy in the studied application domain.

Conclusions:

• The WC and WSPDD algorithms effectively address big data challenges in integrated CBR-BN systems.
• The proposed methods enhance computational efficiency and predictive accuracy for intelligent engineering applications.
• This research contributes to the advancement of big data analytics in complex engineering fields.