Quantifying the interactions among food, energy, and water (FEW) systems is crucial to support integrated policies for the nexus governance. Metropolitan areas are the main consumption and distribution centers of these three resources and, as urbanization continues, their role will become even more central. Despite this, the current understanding of FEW systems in metropolitan regions is limited. In this dissertation, the key factors leading to a more sustainable FEW system are identified in the metropolitan area of Phoenix, Arizona using the integrated WEAP-MABIA-LEAP platform. In this region, the FEW nexus is challenged by dramatic population growth, competition among increasing FEW demand, and limited water availability that could further decrease under climate change. First, it was shown that the WEAP platform allows the reliable simulations of water allocations from supply sources to demand sectors and that agriculture is a key stressor of the nexus, which will require additional groundwater (+83%) and energy (+15%) if cropland area is preserved over the next 50 years. Second, the climate change impacts on the food-water nexus were quantified by applying the WEAP-MABIA model with climate projections up to 2100 from 27 GCMs under different warming levels. It was found that the increases in temperature will lead to higher atmospheric evaporation demand that will, in turn, reduce crop production at a rate of -4.8% per decade. In the last part, the fully integrated WEAP-MABIA-LEAP platform was applied to investigate future scenarios of the FEW nexus in the metropolitan region. Several scenarios targeting each FEW sector were compared through sustainability indicators quantifying availability/consumption, reliability, and productivity of the three resources. Results showed that increasing renewable energy and changing cropping patterns will increase the FEW nexus sustainability compared to business-as-usual conditions. The findings of this dissertation, along with its analytical approach, support policy making towards integrated FEW governance and sustainable development.

Random forest (RF) is a popular and powerful technique nowadays. It can be used for classification, regression and unsupervised clustering. In its original form introduced by Leo Breiman, RF is used as a predictive model to generate predictions for new observations. Recent researches have proposed several methods based on RF for feature selection and for generating prediction intervals. However, they are limited in their applicability and accuracy. In this dissertation, RF is applied to build a predictive model for a complex dataset, and used as the basis for two novel methods for biomarker discovery and generating prediction interval.

Firstly, a biodosimetry is developed using RF to determine absorbed radiation dose from gene expression measured from blood samples of potentially exposed individuals. To improve the prediction accuracy of the biodosimetry, day-specific models were built to deal with day interaction effect and a technique of nested modeling was proposed. The nested models can fit this complex data of large variability and non-linear relationships.

Secondly, a panel of biomarkers was selected using a data-driven feature selection method as well as handpick, considering prior knowledge and other constraints. To incorporate domain knowledge, a method called Know-GRRF was developed based on guided regularized RF. This method can incorporate domain knowledge as a penalized term to regulate selection of candidate features in RF. It adds more flexibility to data-driven feature selection and can improve the interpretability of models. Know-GRRF showed significant improvement in cross-species prediction when cross-species correlation was used to guide selection of biomarkers. The method can also compete with existing methods using intrinsic data characteristics as alternative of domain knowledge in simulated datasets.

Lastly, a novel non-parametric method, RFerr, was developed to generate prediction interval using RF regression. This method is widely applicable to any predictive models and was shown to have better coverage and precision than existing methods on the real-world radiation dataset, as well as benchmark and simulated datasets.

Researchers have found inconsistent effects (negative or positive) of social relationships on self-control capacity. The variation of findings may depend on the aspects of social relationships. In this study, rather than examining overall social relationships and self-control, characteristics in social relationships were clearly defined, including social support, social connection and social conflict, to determine their specific effects on self-control. An online survey study was conducted, and 292 college students filled out the survey. For data analysis, path analysis was utilized to examined the direct effect and indirect effect from social relationships to self-control. Results showed social connection and social conflict may indirectly associate with self-control through stress, but social support does not. It may suggest, in traditional stress buffering model, it is the social connection in social support that really reduce the stress. Concerning the direct effects, social support and social connection were significantly associated with self-control directly, but social conflict does not. This result may support the Social Baseline Theory that positive social relationships have direct regulating effects. Results are good for guidance of experimental manipulation of social relationships in study of social influences of self-control.