A consistency-specificity trade-off to select source behavior in information fusion

Area of Science:

• Artificial Intelligence
• Information Fusion
• Belief Function Theory

Background:

• Combining information from multiple sources with unknown reliability is a significant challenge in belief function theory.
• Existing fusion schemes often lack transparency regarding assumptions about data sources.

Purpose of the Study:

• To propose a new method for selecting source behaviors within a general fusion scheme.
• To enhance the clarity of assumptions made about data sources during information fusion.

Main Methods:

• Developed a fusion scheme that explicitly states assumptions about data sources.
• Introduced and adapted the concepts of specificity and consistency for knowledge representation selection.
• Applied the proposed approach to various examples to demonstrate its efficacy.

Main Results:

• The proposed method effectively selects source behaviors based on specificity and consistency.
• The fusion scheme clarifies the underlying assumptions about the sources.
• Demonstrated that the approach generalizes existing fusion strategies.

Conclusions:

• The novel approach offers a more transparent and adaptable method for information fusion in belief function theory.
• The concepts of specificity and consistency are crucial for reliable source selection.
• This work provides a unified framework that encompasses various existing fusion techniques.