Bayesian Analysis of Target Detection with Enhanced Receiver Operating Characteristic

TL;DR

This paper introduces a Bayesian framework for radar target detection that enhances the traditional ROC curve with a predictive value, aiding operators in threshold decisions.

Contribution

It develops a Bayesian approach incorporating predictive value into ROC analysis for improved radar target detection decision-making.

Findings

Enhanced ROC curve with predictive value for better decision support

Bayesian method accurately predicts target presence or absence

Provides a more informed threshold selection process

Abstract

In the present paper we develop a Bayesian analysis of radar target detection that uses the parameters of conventional radar analysis to provide a valid prediction of target presence or absence when received signals cross or fail to cross chosen threshold values. A Positive Predictive Value parameter is added to the normal Receiver Operating Characteristic to provide information that allows the radar operator to make an informed decision in the choice of threshold.