C.E.Metz, Basic principles of ROC analysis, Seminars Nucl. Med., 8(1978) 283298.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....an optimization problem where the quantity to be maximized is the performance of the classifier on an independent dataset. There are, however, numerous problems with representing classifier performance by a single (scalar) objective function, which is needed so that one can use a scalar optimizer [6], 7] Binary classifiers [4] have, in essence, two implicit objective functions: one describing how well they classify the abnormal cases (sensitivity) and one describing how well they classify the normal cases (specificity) These two objective functions are noncommensurable, implying that it ....

....the fraction of class observations that are correctly classified is used as an estimate of . Likewise, the fraction of class observations that are correctly classified is used as an estimate of . A popular construct used for describing the performance of a diagnostic classifier is the ROC curve [6], 7] 20] 21] A ROC curve is generated by varying the value of one (or more) of the components of the parameter vector , and plotting the corresponding and values. For example, the Fig. 2. The two ROC curves have equal Az values, but, depending upon the relative preferences concerning the ....

[Article contains additional citation context not shown here]

C. E. Metz, "Basic principles of ROC analysis," Seminars Nucl. Med., vol. VIII, no. 4, pp. 283--298, 1978.

....and Swets [33] Provost and Fawcett [34] call this a whole curve metric to di erentiate it from metrics whichevaluate a single point on the ROC curve like those discussed in Section 2.1. The area under the ROC curve has been applied to mathematical psychology [35,36] diagnostic medical imaging [37,38], and more recently to machine learning [39] The area (14) is equivalent to the average power of the most powerful test under a uniform prior on the user s false alarm constraint. AUC### (#) is also equivalent to the probability of error of a Mann Whitney or Wilcoxon rank order test for randomly ....

C. E. Metz, \Basic principles of ROC analysis," Seminars in Nuclear Medicine,vol. 8, no. 4, pp. 283-298, 1978. 40

....needed to optimize any learning model parameters. The idea is to optimize the model parameters without using the test set since using the test set will likely overfit the learner to the test data [10] We used maximization of area under the receiver operating curve (ROC) for parameter optimisation [11]. Thus, each fold has a train, validation, and test set with the proportions of each class in each set maintained. Each set in the fold is further processed as described in this algorithm: For each article in this set Extract mesh terms precede all terms with mh replace all punctuation with ....

Metz, C.E., Basic Principles of ROC Analysis. Seminars in Nuclear Medicine, 1978. 8(4): p. 283298.

....maximal values. However, in this paper we do not have an accurate R a for our data sets (although it is still very interesting to compare R a with R r ) hence we omit the derivation of the theoretical SDR and SNR. In addition, we can perform Receiver Operating Characteristic (ROC) curve analysis [27 28], which involves computing the True Positive Rate (TPR) and the False Positive Rate (FPR) while varying algorithm parameters, such as the BT threshold. TPR indicates the sensitivity of the method, and the False Negative Rate (FNR=1 FPR) indicates the specificity of the method. We can write TPR and ....

Metz, C.E. "Basic principles of ROC analysis," Seminars In Nuclear Medicine, vol.8, pp.283-98, 1978

....difficulties for the J coefficient as well because it requires the continuous probability prediction to be dichotomized. 3.2.2. 5 Receiver Operating Characteristic (ROC) Curves A general solution to the arbitrary thresholds problem mentioned above is Receiver Operating Characteristic (ROC) curves [47]. One selects many cutoff points, from 0 to 1 in our case, and calculates the sensitivity and specificity for each cutoff value, and plots sensitivity against 1 specificity as shown in Figure 4. Such a curve describes the compromises that can be made between sensitivity and specificity as the ....

C. Metz: "Basic Principles of ROC Analysis". In Seminars in Nuclear Medicine, VIII(4):283-298, 1978.

....the image content. Although not dealing with progressive compression, a few previous studies have been close in spirit to the work described in this paper; they compare compression algorithms by an objective recognition task in a reasonably realistic simulation of image use. In [12] 13] 14] [15], still image compression algorithms are evaluated for diagnostic utility by simulating their clinical use by radiologists. In [16] compressed video clips of American Sign Language were compared by deaf subjects for intelligibility. In our study of recognition times for progressive compression ....

C. E. Metz. Basic principles of ROC analysis. Seminars in Nuclear Medicine, VIII(4):282-298, Oct. 1978.

....and 0.75 [18] In fact, and as noted by some authors [92] the choice of cutoff value is arbitrary, and one can obtain different results by selecting different cutoff values. A general solution to the arbitrary thresholds problem mentioned above is Receiver Operating Characteristic (ROC) curves [89]. One selects many cutoff points, from 0 to 1 in our case, and calculates the sensitivity and specificity for each cutoff value, and plots sensitivity against 1 specificity as shown in Figure 10. Such a curve describes the compromises that can be made between sensitivity and specificity as the ....

C. Metz, "Basic Principles of ROC Analysis," Seminars in Nuclear Medicine, vol. VIII, no. 4, pp. 283-298, 1978.

.... scores on quality (e.g. analysis of variance (ANOVA) and receiver operating characteristic (ROC) curves) Examples of such approaches may be found in [31, 16, 32, 33, 15, 14, 34] ROC analysis is the dominant technique for evaluating the suitability of radiologic techniques for real applications [35, 36, 37, 38]. Its origins are in the theory of signal detection: a filtered version of signal plus Gaussian noise is sampled and compared to a threshold. If the threshold is exceeded, then the signal is said to be there. As the threshold varies, the probability of erroneously declaring a signal absent and the ....

C. E. Metz, "Basic principles of ROC analysis," Seminars in Nuclear Medicine, vol. VIII, pp. 282--298, Oct. 1978.

.... scores on quality (e.g. analysis of variance (ANOVA) and receiver operating characteristic (ROC) curves) Examples of such approaches may be found in [31, 16, 32, 33, 15, 14, 34] ROC analysis is the dominant technique for evaluating the suitability of radiologic techniques for real applications [35, 36, 37, 38]. Its origins are in the theory of signal detection: a filtered version of signal plus Gaussian noise is sampled and compared to a threshold. If the threshold is exceeded, then the signal is said to be there. As the threshold varies, the probability of erroneously declaring a signal absent and the ....

C. E. Metz, "Basic principles of ROC analysis," Seminars in Nuclear Medicine, vol. VIII, pp. 282--298, Oct. 1978.

.... of viewers scores on quality (e.g. analysis of variance (ANOVA) and receiver operating characteristic (ROC) curves) Examples of such approaches may be found in [13 15, 30 33] ROC analysis is the dominant technique for evaluating the suitability of radiologic techniques for real applications [34 37]. Its origins are in the theory of signal detection: a filtered version of signal plus Gaussian noise is sampled and compared to a threshold. If the threshold is exceeded, then the signal is said to be there. As the threshold varies, the probability of erroneously declaring a signal absent and the ....

C. E. Metz, "Basic principles of ROC analysis," Seminars in Nuclear Medicine, vol. VIII, pp. 282--298, Oct. 1978.

....can be made by using a DM, classifying a case as a control, or classifying a control as a case. In biostatistical literature the probability that a DM classifies a case as a control is called false negative ratio and probability that a DM classifies a control as a case is false positive ratio (Metz, 1978). Intuitive definitions for true negative ratio and true positive ratio follow. However, often the terms specificity and sensitivity are used. Specificity is 2 the probability that a control is correctly identified by a DM and sensitivity is the probability of correctly classifying a case by a ....

....(1950) introduced operating characteristic (OC) to the statistical decision theory. ROC, inspired by the OC idea, was developed in the context of signal detection theory (Green and Swets, 1966) The medical arena has benefited and used ROC analysis extensively since early 1980 s. Swets (1979) and Metz (1978,1986) pioneered the application of ROC methodology in medical and radiologic imaging. Recently, applications of ROC have been seen in a variety of branches of medicine. A generalization of interest for us is to define a statistic for comparing the sensitivities across a range of all ....

Metz, C.E. (1978). Basic principles of ROC analysis. Seminars in Nuclear Medicine, Vol.VII, No.4, 283-298.

....is most e#ectively displayed using an Table 6. Grouped Data on Number of Nevi Frequencies Number of nevi Controls Cases Relative risks 0 86 44 1.0 1 10 237 215 1.8 11 30 94 188 3.9 31 50 20 60 5.9 50 16 40 4. 9 NOTE: Groupings used are those of Berwick et al. 1996) ROC curve (Metz 1978). The lower solid curve in Figure 1 corresponds to discrimination on the basis of number of nevi alone, whereas the upper curve is based on the logistic discriminant score encompassing the full set of measured risk factors. These curves are constructed in the following way. The horizontal axis ....

Metz, C. E. (1978), "Basic Principles of ROC Analysis," Seminars in Nuclear Medicine, 8, 238--298.

.... and as noted by some authors [51] the choice of cutoff value is arbitrary, and one can obtain different results by selecting different cutoff values (see the example in [28] A general solution to the arbitrary thresholds problem mentioned above is Receiver Operating Characteristic (ROC) curves [49]. One selects many cutoff points, from 0 to 1 in our case, and calculates the sensitivity and specificity for each cutoff value, and plots sensitivity against 1 specificity as shown in Figure 3. Such a curve describes the compromises that can be made between sensitivity and specificity as the ....

C. Metz: "Basic Principles of ROC Analysis". In Seminars in Nuclear Medicine, VIII(4):283-298, 1978.

....A ngerprint matching system can be tuned (by setting an appropriate threshold on the matching score) to operate at a desired value of FAR. However, if we try to decrease the FAR of the system, then it would increase the FRR and vice versa. The Receiver Operating Characteristic (ROC) Curve [107] is a plot of FAR versus FRR which permits the system designer to assess the performance of the recognition system at various operating points (thresholds in the decision rule) In this sense, ROC provides a more comprehensive performance measure than, say, the equal error rate of the system ....

C.E. Metz, \Basic Principles of ROC Analysis," Seminars in Nuclear Medicine, Vol. VIII, No. 4, pp. 283-298, 1978.

....their feature vectors can be used to train a classification system. Ideal candidates for these include neural networks, nearest neighbour classifier and decision trees. The ability of the system to find the correct features and generalise during classification is measured using ROC curves (see Metz, 1978). Figure 1 near here In this paper we focus on ROI detection and single image decomposition into ROI using edge detection methods and fuzzy clustering technique. The paper is organised as follows. In section 2, we present the standard histogram equation method of image enhancement alongside ....

Metz, C., 1978. Basic principles of ROC analysis, Seminars in Nuclear Medicine, 8(4), 283-298.

....MammoNet, has the states: 1 2, 1 3, 2 3, 3 4, and NA. Tabar describes calcification density as varying. We mapped our state 1 3 to Tabar s varying, reasoning that the state 1 3 ranges the most from finest to coarsest density. The performance of the model was assessed using the LABROC1 program [18] for receiver operating characteristic (ROC) analysis (Figure 3) MammoNet s performance as measured by the very high A z value of 0.881 compares very favorably with that of artificial neural network models [28] and expert mammographers. 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 True positive ....

Metz CE. Basic principles of ROC analysis. Seminars in Nuclear Medicine 1978; 4: 283-298.

No context found.

C.E.Metz, Basic principles of ROC analysis, Seminars Nucl. Med., 8(1978) 283298.

No context found.

Metz C.E.: Basic principles of ROC analysis, Semin Nucl Med, 8(4):283-98, 1978.

No context found.

C. E. Metz, "Basic principles of roc analysis," Seminars in Nuclear Medicine, vol. 8, pp. 283--298, 1978.

No context found.

C. E Metz, Basic principles of ROC analysis, Seminars in Nuclear Medicine, vol. 8, issue 4, pp. 283-298, 1978.

No context found.

C. E. Metz, "Basic principles of ROC analysis," Sem. Nucl. Med., vol. 8, pp. 283--298, 1978.

No context found.

C.E. Metz. Basic principles of ROC analysis. Seminars in Nuclear Medicine, VIII(4), October 1978.

No context found.

C. E. Metz. Basic principle of ROC analysis. Seminars Nucl. Med, 8:283-298, 1978.

No context found.

Metz CE. Basic principles of ROC analysis. Semin Nucl Med 1978;8:283--98.

No context found.

Metz CE. Basic principles of ROC analysis. Semin Nucl Med 1978; 8: 283 298.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC