Related Experiment Videos
Likelihood-ratio decision strategy for independent observations in the same-different task: an approximation to the
1Department of Psychology, University of Auckland, New Zealand. j.irwin@auckland.ac.nz
Perception & Psychophysics
|February 1, 1997
Summary
A new approximation simplifies calculating the receiver operating characteristic (ROC) curves for likelihood-ratio decision strategies. This method provides accurate results without complex double integrals, improving decision analysis efficiency.
Area of Science:
- Decision analysis
- Statistical modeling
Background:
- Optimal decision-making relies on the likelihood-ratio criterion.
- Calculating receiver operating characteristic (ROC) curves for this strategy is complex, involving double integrals.
Purpose of the Study:
- To develop a simplified approximation for the parametric equations of ROC curves based on the likelihood-ratio strategy.
- To provide a computationally efficient method for ROC analysis.
Main Methods:
- Developed an approximation to the parametric equations for the likelihood-ratio strategy.
- The approximation involves a pair of equations.
Main Results:
- The approximation yields ROC curves that are virtually indistinguishable from those derived from the full, complex model.
- The simplified equations avoid the need for numerical evaluation of double integrals.
Conclusions:
- The developed approximation offers a practical and accurate method for generating ROC curves.
- This facilitates the application of likelihood-ratio strategies in decision analysis and signal detection theory.