Smooth semiparametric receiver operating characteristic curves for continuous diagnostic tests

Area of Science:

• Statistics
• Biostatistics
• Machine Learning

Background:

• Receiver Operating Characteristic (ROC) curve analysis is crucial for evaluating diagnostic tests.
• Existing non-parametric ROC curve estimators can suffer from bias and inefficiency.
• Smooth non-parametric methods have been proposed but may not be optimal.

Purpose of the Study:

• To develop a novel semiparametric kernel distribution function estimator.
• To construct a new smooth semiparametric estimator for the ROC curve.
• To evaluate the efficiency and performance of the proposed estimators.

Main Methods:

• Semiparametric kernel estimation for distribution functions.
• Construction of a smooth semiparametric ROC curve estimator.
• Derivation of asymptotic bias and variance for proposed estimators.
• Development of data-based bandwidth selection methods.
• Comparative analysis using simulation studies and real data sets.

Main Results:

• The proposed semiparametric distribution function estimator is more efficient than traditional non-parametric kernel estimators.
• The new semiparametric ROC curve estimator outperforms existing smooth non-parametric estimators in terms of bias, standard error, and mean-square error.
• Simulation studies confirm the superiority of the proposed estimators.

Conclusions:

• The novel semiparametric kernel approach offers a more efficient and accurate method for estimating distribution functions and ROC curves.
• The proposed methods provide a valuable alternative for statistical analysis in various fields, including medical diagnostics.
• Data-based bandwidth selection enhances the practical applicability of the estimators.