The combination of neural estimates in prediction and decision problems

In this dissertation, different ways of combining neural predictive models or neural-based forecasts are discussed. The proposed approaches consider mostly Gaussian radial basis function networks, which can be efficiently identified and estimated through recursive/adaptive methods. Two different ways of combining are explored to get a final estimate – model mixing and model synthesis –, with the aim of obtaining improvements both in terms of efficiency and effectiveness. In the context of model mixing, the usual framework for linearly combining estimates from different models is extended, to deal with the case where the forecast errors from those models are correlated. In the context of model synthesis, and to address the problems raised by heavily nonstationary time series, we propose hybrid dynamic models for more advanced time series forecasting, composed of a dynamic trend regressive model (or, even, a dynamic harmonic regressive model), and a Gaussian radial basis function network. Additionally, using the model mixing procedure, two approaches for decision-making from forecasting models are discussed and compared: either inferring decisions from combined predictive estimates, or combining prescriptive solutions derived from different forecasting models. Finally, the application of some of the models and methods proposed previously is illustrated with two case studies, based on time series from finance and from tourism.

Motivating sustainable goal choices and providing effective feedback

Overconsumption of natural resources and the associated environmental hazards are one of today’s most pressing global issues. In the western world, individual consumption in homes and workplaces is a key contributor to this problem. Reflecting the importance of individual action in this domain, this thesis focuses on studying and influencing choices related to sustainability and energy consumption made by people in their daily lives. There are three main components to this work. Firstly, this thesis asserts that people frequently make ineffective consumption reduction goal choices and attempts to understand the rationale for these poor choices by fitting them to goalsetting theory, an established theoretical model of behavior change. Secondly, it presents two approaches that attempt to influence goal choice towards more effective targets, one of which deals with mechanisms for goal priming and the other of which explores the idea that carefully designed toys can exert influence on children’s long term consumption behavior patterns. The final section of this thesis deals with the design of feedback to support the performance of environmentally sound activities. Key contributions surrounding goals include the finding that people choose easy sustainable goals despite immediate feedback as to their ineffectiveness and the discussion and study of goal priming mechanisms that can influence this choice process. Contributions within the design of value instilling toys include a theoretically grounded framework for the design of such toys and a completed and tested prototype toy. Finally, contributions in designing effective and engaging energy consumption feedback include the finding that negative feedback is best presented verbally compared with visually and this is exemplified and presented within a working feedback system. The discussions, concepts, prototypes and empirical findings presented in this work will be useful for both environmental psychologists and for HCI researchers studying eco-feedback.