Fuzzy-Receiver Operating Characteristics (F-ROC) for Fuzzy-Inspired Biosensing Performance Evaluation

In this paper, a novel fuzzy-receiver operating characteristics (F-ROC) is proposed to evaluate the performance of fuzzy-inspired biosensing (FIB). The development process of diseases is fuzzy, and the traditional classification of diseases is an either-or hard classification, which cannot reflect the developmental stages of diseases correctly. FIB utilizes fuzzy theory to realize the disease classification, and the traditional ROC curve is not suitable as an evaluation mechanism for the disease fuzzy process. Therefore, this paper proposes a general model to describe the transfer process of the membership function of FIB, and proposes the reasons why the traditional ROC is not applicable. After that, this paper rewrites the parameter definitions in the ROC curve and establishes the F-ROC curve to evaluate the transfer performance of FIB. An example of tumor stage classification using multi-contrast-agent strategies (MCAS) strategy is utilized to verify the evaluation of the FIB performance of the proposed F-ROC curve.