This thesis describes research into the score-level fusion process in multimodal biometrics. The emphasis of the research is on the fusion of face and voice biometrics in the two recognition modes of verification and open-set identification. The growing interest in the use of multiple modalities in biometrics is due to its potential capabilities for eradicating certain important limitations of unimodal biometrics. One of the factors important to the accuracy of a multimodal biometric system is the choice of the technique deployed for data fusion. To address this issue, investigations are carried out into the relative performance of several statistical data fusion techniques for combining the score information in both unimodal and multimodal biometrics (i.e. speaker and/ or face verification). Another important issue associated with any multimodal technique is that of variations in the biometric data. Such variations are reflected in the corresponding biometric scores, and can thereby adversely influence the overall effectiveness of multimodal biometric recognition. To address this problem, different methods are proposed and investigated. The first approach is based on estimating the relative quality aspects of the test scores and then passing them on into the fusion process either as features or weights. The approach provides the possibility of tackling the data variations based on adjusting the weights for each of the modalities involved according to its relative quality. Another approach considered for tackling the effects of data variations is based on the use of score normalisation mechanisms. Whilst score normalisation has been widely used in voice biometrics, its effectiveness in other biometrics has not been previously investigated. This method is shown to considerably improve the accuracy of multimodal biometrics by appropriately correcting the scores from degraded modalities prior to the fusion process. The investigations in this work are also extended to the combination of score normalisation with relative quality estimation. The experimental results show that, such a combination is more effective than the use of only one of these techniques with the fusion process. The thesis presents a thorough description of the research undertaken, details the experimental results and provides a comprehensive analysis of them.

In this thesis we develop methods for dealing with missing data in a univariate response variable when estimating regression parameters. Missing outcome data is a problem in a number of applications, one of which is follow-up studies. In follow-up studies data is collected at two (or more) occasions, and it is common that only some of the initial participants return at the second occasion. This is the case in Paper II, where we investigate predictors of decline in self reported health in older populations in Sweden, the Netherlands and Italy. In that study, around 50% of the study participants drop out. It is common that researchers rely on the assumption that the missingness is independent of the outcome given some observed covariates. This assumption is called data missing at random (MAR) or ignorable missingness mechanism. However, MAR cannot be tested from the data, and if it does not hold, the estimators based on this assumption are biased. In the study of Paper II, we suspect that some of the individuals drop out due to bad health. If this is the case the data is not MAR. One alternative to MAR, which we pursue, is to incorporate the uncertainty due to missing data into interval estimates instead of point estimates and uncertainty intervals instead of confidence intervals. An uncertainty interval is the analog of a confidence interval but wider due to a relaxation of assumptions on the missing data. These intervals can be used to visualize the consequences deviations from MAR have on the conclusions of the study. That is, they can be used to perform a sensitivity analysis of MAR. The thesis covers different types of linear regression. In Paper I and III we have a continuous outcome, in Paper II a binary outcome, and in Paper IV we allow for mixed effects with a continuous outcome. In Paper III we estimate the effect of a treatment, which can be seen as an example of missing outcome data.