Lae-Jeong Park and Jung-Ho Moon A Learning Method of Directly Optimizing Classifier Performance at Local Operating Range
نویسندگان
چکیده
This paper addresses an effective learning method that enables us to directly optimize neural network classifier's discrimination performance at a desired local operating range by maximizing a partial area under a receiver operating characteristic (ROC) or domain-specific curve, which is difficult to achieve with classification accuracy or mean squared error (MSE)-based learning methods. The effectiveness of the proposed approach is demonstrated in terms of fraud detection capability in the credit card fraud detection, compared with the MSE-based approach. Keyword: Classification, Receiver Operating Characteristic, And Area Under Curve.
منابع مشابه
Accepted Version Efficient Pedestrian Detection by Directly Optimizing the Partial Area under the Roc Curve *
Many typical applications of object detection operate within a prescribed false-positive range. In this situation the performance of a detector should be assessed on the basis of the area under the ROC curve over that range, rather than over the full curve, as the performance outside the range is irrelevant. This measure is labelled as the partial area under the ROC curve (pAUC). Effective casc...
متن کاملEfficient pedestrian detection by directly optimize the partial area under the ROC curve
Many typical applications of object detection operate within a prescribed false-positive range. In this situation the performance of a detector should be assessed on the basis of the area under the ROC curve over that range, rather than over the full curve, as the performance outside the range is irrelevant. This measure is labelled as the partial area under the ROC curve (pAUC). Effective casc...
متن کاملErratum: ACKNOWLEDGMENTS
[This corrects the article on p. 504 in vol. 35, PMID: 22111042.].
متن کاملAnomaly Detection Using SVM as Classifier and Decision Tree for Optimizing Feature Vectors
Abstract- With the advancement and development of computer network technologies, the way for intruders has become smoother; therefore, to detect threats and attacks, the importance of intrusion detection systems (IDS) as one of the key elements of security is increasing. One of the challenges of intrusion detection systems is managing of the large amount of network traffic features. Removing un...
متن کاملEverolimus-Eluting Versus Sirolimus-Eluting Stents in Patients Undergoing Percutaneous Coronary Intervention
Kyung Woo Park, MD,* In-Ho Chae, MD,† Do-Sun Lim, MD,‡ Kyoo-Rok Han, MD,§ Han-Mo Yang, MD,* Hae-Young Lee, MD,* Hyun-Jae Kang, MD,* Bon-Kwon Koo, MD,* Taehoon Ahn, MD, Jung-Han Yoon, MD,¶ Myung-Ho Jeong, MD,# Taek-Jong Hong, MD,** Woo-Young Chung, MD,†† Sang-Ho Jo, MD,‡‡ Young-Jin Choi, MD,‡‡ Seung-Ho Hur, MD,§§ Hyuck-Moon Kwon, MD, Dong-Woon Jeon, MD,¶¶ Byung-Ok Kim, MD,## Si-Hoon Park, MD,***...
متن کامل