Developing Detection DecisioNS on the Absence or Presence of a Radiological Source using a Bayesian Interaction Model

Area of Science:

• Health Physics
• Statistical Modeling
• Radiation Detection

Background:

• Limited research exists on Bayesian decision thresholds for gross count measurements.
• Bayesian modeling offers a structured approach to information processing and statistical inference.

Purpose of the Study:

• To develop and validate a Bayesian interaction model for analyzing gross count measurements.
• To establish a decision threshold for detecting sources using Bayesian statistics.

Main Methods:

• Developed a Bayesian linear regression model analyzing gross counts and their standard deviations over five sequential measurements.
• Conditioned the analysis on whether data originated from background or source measurements.
• Utilized a constructed parameter ζ with a probability distribution for detection decisions.

Main Results:

• The Bayesian model was statistically validated and performed optimally for detecting weaker sources.
• The analysis effectively used sequential data from continuous gross count measurements.
• The parameter ζ provided a statistical measure for detection decisions.

Conclusions:

• The developed Bayesian model offers a promising approach for operational health physics applications.
• The model's advantages include not requiring established training datasets or extensive background measurements.
• The model and parameter ζ are universally applicable beyond the presented predictor variable.