none

Diagnostic tests are central in the field of modern medicine. One of the main factors for interpreting a diagnostic test is the discriminatory accuracy. For a continuous-scale diagnostic test, the area under the receiver operating characteristic (ROC) curve, AUC, is a useful one-number summary index for the diagnostic accuracy of the test. When only a particular region of the ROC curve would be of interest, the partial AUC (pAUC) is a more appropriate index for the diagnostic accuracy. In this thesis, we develop seven confidence intervals for the pAUC under the semi-parametric models for the diseased and non-diseased populations by using the normal approximation, bootstrap and empirical likelihood methods. In addition, we conduct simulation studies to compare the finite sample performance of the proposed confidence intervals for the pAUC. A real example is also used to illustrate the application of the recommended intervals.

No description available.

Fourier Transform Infrared (FTIR) has been use to discriminate different pathogens by signals from cells infected with these versus normal cells as references. To do the statistical analysis, Partial Least Square Regression (PLSR) was utilized to distinguish any two kinds of virus‐infected cells and normal cells. Validation using Bootstrap method and Cross‐validations were employed to calculate the shrinkages of Area Under the ROC Curve (AUC) and specificities corresponding to 80%, 90%, and 95% sensitivities. The result shows that our procedure can significantly discriminate these pathogens when we compare infected cells with the normal cells. On the height of this success, PLSR was applied again to simultaneously compare two kinds of virus‐infected cells and the normal cells. The shrinkage of Volume Under the Surface (VUS) was calculated to do the evaluation of model diagnostic performance. The high value of VUS demonstrates that our method can effectively differentiate virus‐infected cells and normal cells.

Made available in DSpace on 2014-10-09T12:37:02Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T13:56:31Z (GMT). No. of bitstreams: 1 04495.pdf: 1601893 bytes, checksum: e55704a48b48bc206869a88930e54240 (MD5) / Dissertacao (Mestrado) / IPEN/D / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

Properly choosen input data, their form, way of the collection an the following correction as well, are the key factors affecting the accuracy of the more and more popular models of the species model distribution (SDM). A design unfluence of the data to the choosen distribution model of the virtual species has been tested. The General Linear Model (GLM) has been used. Four types of the present-absent collection designs have been tested for a virtual species on the area of the Czech Republic as follows: accidental, systematic, points on the easy accessible areas (near the communications) and the points with the higher concentration of the scientists (an area of CHKO).The TSS (True skills statistics), KAPPA and AUC (Area under curve) have been used to compare the accuracy of the prediction of the models .The points being choosen from the easy to access areas and CHKO had worse results in all of the monitored values than the accidental or the systematic choice of the points from the whole area of the Czech Republic.It results that the data collection mode affects the final accuracy of the species distribution models.

Made available in DSpace on 2014-10-09T12:36:09Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T13:59:22Z (GMT). No. of bitstreams: 1 01803.pdf: 1763975 bytes, checksum: 1d9670fc7ad262d61966c6814c757dc7 (MD5) / Dissertacao (Mestrado) / IPEN/D / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

Made available in DSpace on 2014-10-09T12:37:02Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T13:56:31Z (GMT). No. of bitstreams: 1 04495.pdf: 1601893 bytes, checksum: e55704a48b48bc206869a88930e54240 (MD5) / Dissertacao (Mestrado) / IPEN/D / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

Made available in DSpace on 2014-10-09T12:36:09Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T13:59:22Z (GMT). No. of bitstreams: 1 01803.pdf: 1763975 bytes, checksum: 1d9670fc7ad262d61966c6814c757dc7 (MD5) / Dissertacao (Mestrado) / IPEN/D / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

No description available.