• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제5권3호
• /
• pp.263-267
• /
• 2005

Building classifiers for financial real-world classification problems is often plagued by severely overlapping and highly skewed class distribution. New performance measures such as receiver operating characteristic (ROC) curve and area under ROC curve (AUC) have been recently introduced in evaluating and building classifiers for those kind of problems. They are, however, in-effective to evaluation of classifier's discrimination performance in a particular class of the classification problems that interests lie in only a local operating range of the classifier, In this paper, a new method is proposed that enables us to directly improve classifier's discrimination performance at a desired local operating range by defining and optimizing a partial area under ROC curve or domain-specific curve, which is difficult to achieve with conventional classification accuracy based learning methods. The effectiveness of the proposed approach is demonstrated in terms of fraud detection capability in a real-world fraud detection problem compared with the MSE-based approach.

• Communications for Statistical Applications and Methods
• /
• 제27권2호
• /
• pp.211-224
• /
• 2020

Sensitivity and specificity are popular measures described by the receiver operating characteristic (ROC) curve. There are also two other measures such as the positive predictive value (PPV) and negative predictive value (NPV); however, the PPV and NPV cannot be represented by the ROC curve. Based on the total operating characteristic (TOC) curve suggested by Pontius and Si (International Journal of Geographical Information Science, 97, 570-583, 2014), explanatory methods are proposed to geometrically describe the PPV and NPV by the TOC curve. It is found that the PPV can be regarded as the slope of the right-angled triangle connecting the origin to a certain point on the TOC curve, while 1 - NPV can be represented as the slope of the right-angled triangle connecting a certain point to the top right corner of the TOC curve. When the neutral zone exists, the PPV and 1-NPV can be described as the slopes of two other right-angled triangles of the TOC curve. Therefore, both the PPV and NPV can be estimated using the TOC curve, whether or not the neutral zone is present.

• Asian Pacific Journal of Cancer Prevention
• /
• 제15권10호
• /
• pp.4143-4145
• /
• 2014

Background: This study used receiver operating characteristic curve to analyze Surveillance, Epidemiology and End Results (SEER) Ewing sarcoma (ES) outcome data. The aim of this study was to identify and optimize ES-specific survival prediction models and sources of survival disparities. Materials and Methods: This study analyzed socio-economic, staging and treatment factors available in the SEER database for ES. 1844 patients diagnosed between 1973-2009 were used for this study. For the risk modeling, each factor was fitted by a Generalized Linear Model to predict the outcome (bone and joint specific death, yes/no). The area under the receiver operating characteristic curve (ROC) was computed. Similar strata were combined to construct the most parsimonious models. Results: The mean follow up time (S.D.) was 74.48 (89.66) months. 36% of the patients were female. The mean (S.D.) age was 18.7 (12) years. The SEER staging has the highest ROC (S.D.) area of 0.616 (0.032) among the factors tested. We simplified the 4-layered risk levels (local, regional, distant, un-staged) to a simpler non-metastatic (I and II) versus metastatic (III) versus un-staged model. The ROC area (S.D.) of the 3-tiered model was 0.612 (0.008). Several other biologic factors were also predictive of ES-specific survival, but not the socio-economic factors tested here. Conclusions: ROC analysis measured and optimized the performance of ES survival prediction models. Optimized models will provide a more efficient way to stratify patients for clinical trials.

• Asian Pacific Journal of Cancer Prevention
• /
• 제15권16호
• /
• pp.6781-6785
• /
• 2014

Purpose: This study used receiver operating characteristic curves to analyze Surveillance, Epidemiology and End Results (SEER) medulloblastoma (MB) and primitive neuroectodermal tumor (PNET) outcome data. The aim of this study was to identify and optimize predictive outcome models. Materials and Methods: Patients diagnosed from 1973 to 2009 were selected for analysis of socio-economic, staging and treatment factors available in the SEER database for MB and PNET. For the risk modeling, each factor was fitted by a generalized linear model to predict the outcome (brain cancer specific death, yes/no). The area under the receiver operating characteristic curve (ROC) was computed. Similar strata were combined to construct the most parsimonious models. A Monte Carlo algorithm was used to estimate the modeling errors. Results: There were 3,702 patients included in this study. The mean follow up time (S.D.) was 73.7 (86.2) months. Some 40% of the patients were female and the mean (S.D.) age was 16.5 (16.6) years. There were more adult MB/PNET patients listed from SEER data than pediatric and young adult patients. Only 12% of patients were staged. The SEER staging has the highest ROC (S.D.) area of 0.55 (0.05) among the factors tested. We simplified the 3-layered risk levels (local, regional, distant) to a simpler non-metastatic (I and II) versus metastatic (III) model. The ROC area (S.D.) of the 2-tiered model was 0.57 (0.04). Conclusions: ROC analysis optimized the most predictive SEER staging model. The high under staging rate may have prevented patients from selecting definitive radiotherapy after surgery.

• Communications for Statistical Applications and Methods
• /
• 제29권5호
• /
• pp.533-545
• /
• 2022

Collection of data on several variables, especially in the field of medicine, results in the problem of measurement errors. The presence of such measurement errors may influence the outcomes or estimates of the parameter in the model. In classification scenario, the presence of measurement errors will affect the intrinsic cum summary measures of Receiver Operating Characteristic (ROC) curve. In the context of ROC curve, only a few researchers have attempted to study the problem of measurement errors in estimating the area under their respective ROC curves in the framework of univariate setup. In this paper, we work on the estimation of area under the multivariate ROC curve in the presence of measurement errors. The proposed work is supported with a real dataset and simulation studies. Results show that the proposed bias-corrected estimator helps in correcting the AUC with minimum bias and minimum mean square error.

• 한국물환경학회지
• /
• 제33권2호
• /
• pp.130-139
• /
• 2017

With the need to evaluate accuracy of real-time measurement of cyanobacterial fluorescence to determine cyanobacterial blooms, this research examined 357 paired data (2013-2016) comprising both microscopic toxic cyanobacterial cell counts and concurrent real-time cyanobacterial concentrations at 2 sites (YS1 and YS2) in Yeongsan river. The increase in real-time cyanobacterial concentration was closely associated with the exceedance of 5,000 cyanobacterial cells/ml (odds ratio [OR] 1.07, 95% confidence interval [CI] 1.03-1.12) and 10,000 cells/ml (OR 1.08, 95% CI 1.04-1.12) at YS2 site. The area under the receiver operating characteristic (ROC) curve for the real-time cyanobacterial measurement at the YS2 site was 0.93, which indicates the measurement provides a high accurate detection of cyanobacterial blooms. On the ROC curve, the early alert levels of real-time cyanobacteria ranging $16-23{\mu}g$ chl-a/L would produce acceptable sensitivity of 79% and specificities greater than 90%. The real-time fluorescence measurement was found to be an accurate indicator of cyanobacteria and can serve as a tool for detecting toxic cyanobacterial bloom events in Youngsan river.

• ETRI Journal
• /
• 제41권6호
• /
• pp.797-810
• /
• 2019

Image saliency detection is the basis of perceptual image processing, which is significant to subsequent image processing methods. Most saliency detection methods can detect only a single object with a high-contrast background, but they have no effect on the extraction of a salient object from images with complex low-contrast backgrounds. With the prior knowledge, this paper proposes a method for detecting salient objects by combining the boundary contrast map and the geodesics-like maps. This method can highlight the foreground uniformly and extract the salient objects efficiently in images with low-contrast backgrounds. The classical receiver operating characteristics (ROC) curve, which compares the salient map with the ground truth map, does not reflect the human perception. An ROC curve with distance (distance receiver operating characteristic, DROC) is proposed in this paper, which takes the ROC curve closer to the human subjective perception. Experiments on three benchmark datasets and three low-contrast image datasets, with four evaluation methods including DROC, show that on comparing the eight state-of-the-art approaches, the proposed approach performs well.

• Imaging Science in Dentistry
• /
• 제30권3호
• /
• pp.155-158
• /
• 2000

1. 일반방사선사진과 칼라화한 방사선사진의 비교에서 각각 필름에서 진단을 시행할 때 ROC해석법에서는 true positive fraction (TPF), false positive fraction (FPF)를 매개변수로 하고 있으므로 우선 두가지 필름형태에 대해 각각 따로 다음과 같이 평가한다. 2. 판정기준 병변없다 A, 거의 없다 B, 모르겠다 C, 거의 있다 D, 있다 E 먼저 일반방사선사진에서 실제로 병소가 총있는 것이 50, 총없는 것이 50인데 위 판정기준 각각에 대해(equation omitted) 3. 곡선만들기 a.횡축은 FPF 종축은 TPF로 한 그래프를 plot를 한다. sensitivity 17/50 specificity 26/50 accuracy 43/100 b. 곡선만들기 프로그램을 이용하여 곡선을 만들시에는 TPF를 a에 입력하고 PFP를 b에 입력한다. 이 plot을 그릴 수 있는 프로그램은 http://www.members.tripod.co.kr/jdakim 또는 http://www.chosun.ac.kr/∼jdakim의 홈페이지내 공개자료실에서 다운 받으실 수 있습니다. (equation omitted) 이 프로그램에서 입력할 a, b의 값은 (equation omitted) 위와같이 입력하여 얻어진 일반방사선사진에서의 판독 결과 얻어진 곡선이 그래프에서 곡선이 된다. 이와 같은 커브를 컬러화한 사진 판독에서 똑같이 시행하여 ROC곡선(윗곡선)을 만든 다음 두 곡선을 비교하여 아래면적이 더 큰 쪽이 병소 판독에 우수하다고 결론짓는다.

• Asian Pacific Journal of Cancer Prevention
• /
• 제17권1호
• /
• pp.347-352
• /
• 2016

Background: This study used receiver operating characteristic curve to analyze Surveillance, Epidemiology and End Results (SEER) adenosquamous carcinoma data to identify predictive models and potential disparities in outcome. Materials and Methods: This study analyzed socio-economic, staging and treatment factors available in the SEER database for adenosquamous carcinoma. For the risk modeling, each factor was fitted by a generalized linear model to predict the cause specific survival. An area under the receiver operating characteristic curve (ROC) was computed. Similar strata were combined to construct the most parsimonious models. Results: A total of 20,712 patients diagnosed from 1973 to 2009 were included in this study. The mean follow up time (S.D.) was 54.2 (78.4) months. Some 2/3 of the patients were female. The mean (S.D.) age was 63 (13.8) years. SEER stage was the most predictive factor of outcome (ROC area of 0.71). 13.9% of the patients were un-staged and had risk of cause specific death of 61.3% that was higher than the 45.3% risk for the regional disease and lower than the 70.3% for metastatic disease. Sex, site, radiotherapy, and surgery had ROC areas of about 0.55-0.65. Rural residence and race contributed to socioeconomic disparity for treatment outcome. Radiotherapy was underused even with localized and regional stages when the intent was curative. This under use was most pronounced in older patients. Conclusions: Anatomic stage was predictive and useful in treatment selection. Under-staging may have contributed to poor outcome.

• 한국한의학연구원논문집
• /
• 제17권2호
• /
• pp.107-113
• /
• 2011

Objectives : Sasang typology is extensively studied for the Sasang constitution diagnosis objectification with various data, for example, questionaires, reference materials, etc and analyzed with the several statistical methods. In this study, we used ROC-curve (Receiver Operating Characteristic curve) analysis to diagnose Sasang constitution, which is a kind of epidemiologic research methods and is away from traditional statistical methods. Methods : We collected personality questionnaire which consists of 15 items, from 24 oriental medical clinics. We analyzed the sensitivity and specificity using ROC curve method based on the score of personality questionnaire and also investigated classification accuracy and cut-off value of Sasang constitution. Results : The AUC (area under the ROC curve) value was 0.508 (p=.5511) for Taeeumin, 0.629 (p<.0001) for Soeumin and 0.604(p<.0001) for Soyangin, respectively. so the classification accuracy for Soeumin was highest Soeumin for over 30 points and Soyangin for below 28 points respectively. Conclusions : We suggest that Taeeumin is not classified easily in the ROC-curve analysis. We may classify Soeumin and Soyangin but the accuracy of Sasang constitutional diagnosis is still low.