Approximate Bayesian techniques for inference in stochastic dynamical systems

This thesis is concerned with approximate inference in dynamical systems, from a variational Bayesian perspective. When modelling real world dynamical systems, stochastic differential equations appear as a natural choice, mainly because of their ability to model the noise of the system by adding a variant of some stochastic process to the deterministic dynamics. Hence, inference in such processes has drawn much attention. Here two new extended frameworks are derived and presented that are based on basis function expansions and local polynomial approximations of a recently proposed variational Bayesian algorithm. It is shown that the new extensions converge to the original variational algorithm and can be used for state estimation (smoothing). However, the main focus is on estimating the (hyper-) parameters of these systems (i.e. drift parameters and diffusion coefficients). The new methods are numerically validated on a range of different systems which vary in dimensionality and non-linearity. These are the Ornstein-Uhlenbeck process, for which the exact likelihood can be computed analytically, the univariate and highly non-linear, stochastic double well and the multivariate chaotic stochastic Lorenz '63 (3-dimensional model). The algorithms are also applied to the 40 dimensional stochastic Lorenz '96 system. In this investigation these new approaches are compared with a variety of other well known methods such as the ensemble Kalman filter / smoother, a hybrid Monte Carlo sampler, the dual unscented Kalman filter (for jointly estimating the systems states and model parameters) and full weak-constraint 4D-Var. Empirical analysis of their asymptotic behaviour as a function of observation density or length of time window increases is provided.

Foreign direct investment and corruption

In the last few decades, the world has witnessed an enormous growth in the volume of foreign direct investment (FDI). The global stock of FDI reached US$ 7.5 trillion in 2003 and accounted for 11% of world Gross Domestic Product, up from 7% in 1990. The sales of multinational enterprises at around US$ 19 trillion were more than double the level of world exports. Substantial FDI inflows went into transition countries. Inflows into one of the region's largest recipient, the Russian Federation, almost doubled, enabling Russia to become one of the five top FDI destinations in 2005-2006. FDI inflows in Russia have increased almost threefold from 13.6% in 2003 to 35% in 2007. In 2003, these flows were twice greater than those into China; whilst in 2007 they were six times larger. Russia's FDI inflows were also about 2.5 times greater than those of Brazil. Efficient government institutions are argued by many economists to foster FDI and growth as a result. However, the magnitude of this effect has yet to be measured. This thesis takes a Political Economy approach to explore, empirically, the potential impact of malfunctioning governmental institutions, proxied by three indices of perceived corruption, on FDI stocks accumulation/distribution within Russia over the period of 2002-2004. Using a regional data-set it concentrates on three areas relating to FDI. Firstly, it considers the significance, the size and the sign of the impact of perceived corruption on accumulation of FDI stocks within Russia. Secondly, it quantifies the impact of perceived corruption on the volume of FDI stocks simultaneously estimating the impact of the investment in public capital such as telecommunications and transportation networks on FDI in the presence of corruption. In particular, it addresses the question whether more corrupt regions in Russia are also those that could have accumulated more of FDI stocks, and investigates whether those 'more corrupt' regions would have had lower level of public capital investment. Finally, it examines whether decentralisation increases or decreases corruption and whether a larger extent of decentralisation has a positive or negative impact on FDI (stocks). The results of three studies are as follows. Firstly, along with market potential, corruption is found to be one of the key factors in explaining FDI distribution within Russia between 2002 and 2004. Secondly, corruption on average is found to be related to FDI positively suggesting that it may act as speed money: to save their time foreign direct investors might be willing to bribe the regional authorities so to move in front of the bureaucratic lines. Thirdly, although when corruption is controlled for, the impact of the latter on unobservable FDI is found to be on average positive, no association between FDI and public investment is observed with the only exception of transportation infrastructure (i.e., railway). The results might suggest therefore that it is possible that not only regions with high levels of perceived corruption attract more FDI but also that expansions in public capital investments are not accompanied by an increase of the volume of FDI (stocks) in regions with high levels of corruption. This casts some doubt on the productivity of the investment in public capital in these regions as it might be that bureaucrats may prefer to use these infrastructural projects for rent extraction. Finally, we find decentralisation to have a significant and positive impact on both FDI stock accumulation and corruption, suggesting that local governments may spend more on public goods to make the area more attractive to foreign investors but at the same time they may be interested into extracting rents from foreign investors. These results support the idea that the regulation of FDI is associated with and facilitated by a larger public sector, which distorts competition and introduces opportunities for rent-seeking by particular economic and political factors.

Developments in technologies for human-centric mobile computing and applications

With recent expansions in technology, mobile computing continues to play a vital role in all aspects of our lives. Digital technology tools such as Web browsing, media tracking, social media, and emailing have made mobile technology more than just a means of communication but has widespread use in business and social networks. Developments in Technologies for Human-Centric Mobile Computing and Applications is a comprehensive collection of knowledge and practice in the development of technologies in human –centric mobile technology. This book focuses on the developmental aspects of mobile technology; bringing together researchers, educators, and practitioners to encourage readers to think outside of the box.