臺灣博碩士論文加值系統

在多類別分類問題中，表現機率是分類精確性及分類指標的區辨力的合理評估準則。利用表現機率集合在接收者操作特徵空間中的幾何性質，最大期望效用準則可給出最佳分類程序的明確表達式。同時，透過參數化最佳接收者操作特徵流形，我們可分析其規範性及微分結構。此流形的經驗估計式可被證明具備收斂至一高斯過程，並藉此建立同步信賴區間。延伸二分類中的曲線下面積，此流形下的超體積可做為區辨力的摘要測度。我們給出此超體積可定義以及具備統計意義的條件，並進一步證明其與正確分類機率的等價關係。因此在適當假設下，我們給出對應的U估計式以及參數模型統計推論。相關的數值實驗及資料分析佐證此理論架構的實用性，並可幫助研究者評估多類別分類指標及分類程序。

To evaluate overall discrimination capacity of a marker for multi-class classification tasks, the performance function is a natural assessment tool and fully provides the essential ingredients in receiver operating characteristic (ROC) analysis. The connection between admissible and utility classifiers facilitates illustrating the optimality of likelihood ratio scores as well as constructing a parameterized optimal ROC manifold. The manifolds supply a geometric characterization of the magnitude of separation among multiple classes. It is shown that the hypervolume under the optimal ROC manifold (HUM) is a well-defined and meaningful accuracy measure only in suitable ROC subspaces. Moreover, we provide a rigorous proof for the equality of HUM and its alternative form, the correctness probability, which is directly related to an explicit U-estimator. Our theoretical framework further allows more sophisticated modeling on performance of markers and helps practitioners examine the optimality of applied classification procedures.

Abstract i
中文摘要 ii
1 Introduction 1
1.1 ROC Analysis for Sequential Classification Procedures .......... 2
1.2 Optimal Classification ........................... 3
1.3 Main Achievements ............................. 4
2 Performance of Classifiers 5
2.1 General ROC Space ............................. 5
2.2 Optimality and Utility ............................ 6
2.3 Utility Classifiers in the Decision Space .................. 8
3 Optimal ROC Manifolds 11
3.1 Construction of Manifolds ......................... 11
3.2 Estimation and Inference Procedures for Manifolds ............ 13
4 Hypervolumes under Optimal ROC Manifolds 16
4.1 HUM as an Assessment Measure ...................... 16
4.2 Estimation and Inference Procedures for HUM .............. 18
4.3 Model-based HUM ............................. 21
5 Numerical Experiments and Application 25
5.1 Scenario I: Multinomial Logistic Regression ................ 25
5.2 Scenario II: Multivariate Normal Marker .................. 27
5.3 Scenario III: Univariate Normal Marker .................. 27
5.4 Application to Hepatic Enzyme Profile ................... 29
6 Conclusive Discussion 34
6.1 Limitation of Prediction Probability .................... 34
6.2 Alternative Parameterization .................... 35
6.3 Markers with Discrete or Mixture Distributions .............. 35
6.4 Comparisons among Optimal ROC Manifolds ............... 36
Bibliography 38

Albert, A. and Harris, E. K. (1987). Multivariate Interpretation of Clinical Laboratory Data. CRC Press, first edition.
Beineke, L. W. and Wilson, R. J. (2004). Topics in Algebraic Graph Theory. Cambridge University Press, Cambridge.
Dreiseitl, S., Ohno-Machado, L., and Binder, M. (2000). Comparing three-class diagnostic tests by three-way ROC analysis. Medical Decision Making, 20(3):323–331.
Edwards, D. C., Metz, C. E., and Kupinski, M. A. (2004). Ideal observers and optimal ROC hypersurfaces in n-class classification. IEEE Tansactions on Medical Imaging, 23(7):891–895.
Edwards, D. C., Metz, C. E., and Nishikawa, R. M. (2005). The hypervolume under the ROC hypersurface of “near-guessing” and “near-perfect” observers in n-class classification tasks. IEEE Transactions on Medical Imaging, 24(3):293–299.
He, X., Metz, C., Tsui, B., Links, J., and Frey, E. (2006). Three-class ROC analysis—a decision theoretic approach under the ideal observer framework. IEEE Transactions on Medical Imaging, 25(5):571–581.
Jost, J. (2008). Riemannian Geometry and Geometric Analysis. Springer, New York, fifth edition.
Lesaffre, E. and Albert, A. (1989). Multiple-group logistic regression diagnostics. Journal of the Royal Statistical Society. Series C (Applied Statistics), 38(3):425–440.
Li, J. and Fine, J. P. (2008). ROC analysis with multiple classes and multiple tests: methodology and its application in microarray studies. Biostatistics, 9(3):566 –576.
Lin, D. Y., Wei, L. J., Yang, I., and Ying, Z. (2000). Semiparametric regression for the mean and rate functions of recurrent events. Journal of the Royal Statistical Society. Series B (Statistical Methodology), 62(4):711–730.
Loomis, L. H. and Whitney, H. (1949). An inequality related to the isoperimetric inequality. Bulletin of the American Mathematical Society, 55(10):961–962.
Meyer, M. (1988). A volume inequality concerning sections of convex sets. Bulletin of the London Mathematical Society, 20(2):151 –155.
Mossman, D. (1999). Three-way ROCs. Medical Decision Making, 19(1):78 –89.
Nakas, C. T. and Yiannoutsos, C. T. (2004). Ordered multiple-class ROC analysis with continuous measurements. Statistics in Medicine, 23(22):3437–3449.
Randles, R. H. (1982). On the asymptotic normality of statistics with estimated parameters. The Annals of Statistics, 10(2):462–474.
Schubert, C. M., Thorsen, S. N., and Oxley, M. E. (2011). The ROC manifold for classification systems. Pattern Recognition, 44(2):350–362.
Scurfield, B. K. (1996). Multiple-event forced-choice tasks in the theory of signal detectability. Journal of Mathematical Psychology, 40(3):253–269.
Scurfield, B. K. (1998). Generalization of the theory of signal detectability to n-event m-dimensional forced-choice tasks. Journal of Mathematical Psychology, 42(1):5–31.
van der Vaart, A. W. (1998). Asymptotic Statistics. Cambridge University Press, Cambridge.
van der Vaart, A. W. and Wellner, J. A. (1996). Weak Convergence and Empirical Processes. Springer, New York.