A Study on the Bi-Rayleigh ROC Curve Model
Journal Title: Bonfring International Journal of Data Mining - Year 2012, Vol 2, Issue 2
Abstract
Receiver Operating Characteristic (ROC) curves are used to describe and compare the accuracy of diagnostic test or the ability of a continuous biomarker in discriminating between the subjects into healthy or diseased cases in medical field. The most familiar form of ROC curve is Bi-normal (Gaussian) ROC curve model, which assumes that the test scores or a monotone transformation of the test scores are from two normal populations (i.e. healthy and diseased). It may not be true all the time, it may violate the assumptions of normal distribution in some situations and also we cannot adopt the model as it is when the sample size is small. In this paper, we have proposed ROC curve model for Rayleigh distribution which can be used even when sample size is small. The properties of Bi-Rayleigh ROC model are studied and Area Under the ROC Curve (AUC) are derived. The proposed model is supported by real life example as well as simulation studies. The confidence interval for the population parameter is studied with simulation studies of varying sample sizes. It is found that Bi-Rayleigh ROC model provides better accuracy of classification than the conventional bi-normal ROC model.
Authors and Affiliations
Sudesh Pundir , Amala R
Keywords
Related Articles
- EP ID EP130454
- DOI 10.9756/BIJDM.1358
- Views 113
- Downloads 0
Money Laundering Identification Using Risk and Structural Framework Estimation
Money laundering refers to activities that disguise money receive through illegal operations and make them legitimate. It leaves serious consequence that may lead to economy corruption. One such problem consisting large...
Crop Advisor: A Software Tool for Forecasting Paddy Yield
The highly erratic rainfall and associated climatic parameters in India, have greater influence on the performance of cropping systems and are adversely affecting the crop yields. Forecasting of crop yields from the clim...
RST Approach for Efficient CARs Mining
In data mining, an association rule is a pattern that states the occurrence of two items (premises and consequences) together with certain probability. A class association rule set (CARs) is a subset of association rules...
Encryption Based Hybrid DWT-SVD Watermarking Technique for Data Hiding
This paper widely deals with how data hiding can be more effectively done by encryption based hybrid image watermarking viz., DWT and SWD. Then an image encryption concept is used to protect the image contents. The basic...
An Efficient Meta Scheduling based Virtual Consolidation for Resource Sharing in Green Cloud
In modern researchers, cloud parallel data processing has emerging resource that to be one of the problematic application for Infrastructure-as-a-Service (IaaS) clouds. Major Cloud processing companies include starting i...