The Neural Network ROC Curve Evaluation and Analysis
A neural network ROC curve evaluation is a plot generated by varying a set of trade-off (or threshold) points between the sensitivity and the difference of (1 – specificity) for cases being classified as malignant breast cancer, ranging from 0 to 1. In our simulation modeling in producing the neural network ROC curve evaluation, a set of T values was first adjusted to run discrete units from 0 to 1 in the neural network diagnostic model according to Equation (14) and a set of corresponding sensitivity and (1-specificity) values was calculated when performing the diagnostic model using the clinical instances from the mammographic mass dataset.