Modelling financial time series with switching state space models

The deficiencies of stationary models applied to financial time series are well documented. A special form of non-stationarity, where the underlying generator switches between (approximately) stationary regimes, seems particularly appropriate for financial markets. We use a dynamic switching (modelled by a hidden Markov model) combined with a linear dynamical system in a hybrid switching state space model (SSSM) and discuss the practical details of training such models with a variational EM algorithm due to [Ghahramani and Hilton,1998]. The performance of the SSSM is evaluated on several financial data sets and it is shown to improve on a number of existing benchmark methods.

Urban governance under neoliberalism:New Labour and the restructuring of state-space

In the UK there has been a proliferation of agencies at differing regulatory scales as part of the rescaling and restructuring of the state by New Labour, following the neoliberal policies of previous Conservative governments. This raises questions concerning the extent to which New Labour's urban state restructuring is embedded within neoliberalism, and the local tensions and contradictions arising from emergent New Labour urban state restructuring. This paper examines these questions through the analysis of key policy features of New Labour, and the in-depth exploration of two programmes that are reshaping urban governance arrangements, namely Local Strategic Partnerships (LSPs) and New Deal for Communities (NDC) programmes. We conclude that New Labour's restructuring is best understood in terms of the extended reproduction (roll-out) of neoliberalism. While these “new institutional fixes” are only weakly established and exhibit internal contradictions and tensions, these have not led to a broader contestation of neoliberalism.

A Bayesian state space modelling approach to probabilistic quantitative precipitation forecasting

The generation of very short range forecasts of precipitation in the 0-6 h time window is traditionally referred to as nowcasting. Most existing nowcasting systems essentially extrapolate radar observations in some manner, however, very few systems account for the uncertainties involved. Thus deterministic forecast are produced, which have a limited use when decisions must be made, since they have no measure of confidence or spread of the forecast. This paper develops a Bayesian state space modelling framework for quantitative precipitation nowcasting which is probabilistic from conception. The model treats the observations (radar) as noisy realisations of the underlying true precipitation process, recognising that this process can never be completely known, and thus must be represented probabilistically. In the model presented here the dynamics of the precipitation are dominated by advection, so this is a probabilistic extrapolation forecast. The model is designed in such a way as to minimise the computational burden, while maintaining a full, joint representation of the probability density function of the precipitation process. The update and evolution equations avoid the need to sample, thus only one model needs be run as opposed to the more traditional ensemble route. It is shown that the model works well on both simulated and real data, but that further work is required before the model can be used operationally. © 2004 Elsevier B.V. All rights reserved.

Hidden state models for time series

Amongst all the objectives in the study of time series, uncovering the dynamic law of its generation is probably the most important. When the underlying dynamics are not available, time series modelling consists of developing a model which best explains a sequence of observations. In this thesis, we consider hidden space models for analysing and describing time series. We first provide an introduction to the principal concepts of hidden state models and draw an analogy between hidden Markov models and state space models. Central ideas such as hidden state inference or parameter estimation are reviewed in detail. A key part of multivariate time series analysis is identifying the delay between different variables. We present a novel approach for time delay estimating in a non-stationary environment. The technique makes use of hidden Markov models and we demonstrate its application for estimating a crucial parameter in the oil industry. We then focus on hybrid models that we call dynamical local models. These models combine and generalise hidden Markov models and state space models. Probabilistic inference is unfortunately computationally intractable and we show how to make use of variational techniques for approximating the posterior distribution over the hidden state variables. Experimental simulations on synthetic and real-world data demonstrate the application of dynamical local models for segmenting a time series into regimes and providing predictive distributions.

The 'moving targets' training algorithm

A simple method for training the dynamical behavior of a neural network is derived. It is applicable to any training problem in discrete-time networks with arbitrary feedback. The algorithm resembles back-propagation in that an error function is minimized using a gradient-based method, but the optimization is carried out in the hidden part of state space either instead of, or in addition to weight space. Computational results are presented for some simple dynamical training problems, one of which requires response to a signal 100 time steps in the past.

The 'moving targets' training algorithm

A simple method for training the dynamical behavior of a neural network is derived. It is applicable to any training problem in discrete-time networks with arbitrary feedback. The method resembles back-propagation in that it is a least-squares, gradient-based optimization method, but the optimization is carried out in the hidden part of state space instead of weight space. A straightforward adaptation of this method to feedforward networks offers an alternative to training by conventional back-propagation. Computational results are presented for simple dynamical training problems, with varied success. The failures appear to arise when the method converges to a chaotic attractor. A patch-up for this problem is proposed. The patch-up involves a technique for implementing inequality constraints which may be of interest in its own right.

Pragmatic approach to the development of robust real-time protocols

This research is concerned with the development of distributed real-time systems, in which software is used for the control of concurrent physical processes. These distributed control systems are required to periodically coordinate the operation of several autonomous physical processes, with the property of an atomic action. The implementation of this coordination must be fault-tolerant if the integrity of the system is to be maintained in the presence of processor or communication failures. Commit protocols have been widely used to provide this type of atomicity and ensure consistency in distributed computer systems. The objective of this research is the development of a class of robust commit protocols, applicable to the coordination of distributed real-time control systems. Extended forms of the standard two phase commit protocol, that provides fault-tolerant and real-time behaviour, were developed. Petri nets are used for the design of the distributed controllers, and to embed the commit protocol models within these controller designs. This composition of controller and protocol model allows the analysis of the complete system in a unified manner. A common problem for Petri net based techniques is that of state space explosion, a modular approach to both the design and analysis would help cope with this problem. Although extensions to Petri nets that allow module construction exist, generally the modularisation is restricted to the specification, and analysis must be performed on the (flat) detailed net. The Petri net designs for the type of distributed systems considered in this research are both large and complex. The top down, bottom up and hybrid synthesis techniques that are used to model large systems in Petri nets are considered. A hybrid approach to Petri net design for a restricted class of communicating processes is developed. Designs produced using this hybrid approach are modular and allow re-use of verified modules. In order to use this form of modular analysis, it is necessary to project an equivalent but reduced behaviour on the modules used. These projections conceal events local to modules that are not essential for the purpose of analysis. To generate the external behaviour, each firing sequence of the subnet is replaced by an atomic transition internal to the module, and the firing of these transitions transforms the input and output markings of the module. Thus local events are concealed through the projection of the external behaviour of modules. This hybrid design approach preserves properties of interest, such as boundedness and liveness, while the systematic concealment of local events allows the management of state space. The approach presented in this research is particularly suited to distributed systems, as the underlying communication model is used as the basis for the interconnection of modules in the design procedure. This hybrid approach is applied to Petri net based design and analysis of distributed controllers for two industrial applications that incorporate the robust, real-time commit protocols developed. Temporal Petri nets, which combine Petri nets and temporal logic, are used to capture and verify causal and temporal aspects of the designs in a unified manner.

A design environment for deadlock-free concurrent software

Using current software engineering technology, the robustness required for safety critical software is not assurable. However, different approaches are possible which can help to assure software robustness to some extent. For achieving high reliability software, methods should be adopted which avoid introducing faults (fault avoidance); then testing should be carried out to identify any faults which persist (error removal). Finally, techniques should be used which allow any undetected faults to be tolerated (fault tolerance). The verification of correctness in system design specification and performance analysis of the model, are the basic issues in concurrent systems. In this context, modeling distributed concurrent software is one of the most important activities in the software life cycle, and communication analysis is a primary consideration to achieve reliability and safety. By and large fault avoidance requires human analysis which is error prone; by reducing human involvement in the tedious aspect of modelling and analysis of the software it is hoped that fewer faults will persist into its implementation in the real-time environment. The Occam language supports concurrent programming and is a language where interprocess interaction takes place by communications. This may lead to deadlock due to communication failure. Proper systematic methods must be adopted in the design of concurrent software for distributed computing systems if the communication structure is to be free of pathologies, such as deadlock. The objective of this thesis is to provide a design environment which ensures that processes are free from deadlock. A software tool was designed and used to facilitate the production of fault-tolerant software for distributed concurrent systems. Where Occam is used as a design language then state space methods, such as Petri-nets, can be used in analysis and simulation to determine the dynamic behaviour of the software, and to identify structures which may be prone to deadlock so that they may be eliminated from the design before the program is ever run. This design software tool consists of two parts. One takes an input program and translates it into a mathematical model (Petri-net), which is used for modeling and analysis of the concurrent software. The second part is the Petri-net simulator that takes the translated program as its input and starts simulation to generate the reachability tree. The tree identifies `deadlock potential' which the user can explore further. Finally, the software tool has been applied to a number of Occam programs. Two examples were taken to show how the tool works in the early design phase for fault prevention before the program is ever run.

Mathematical methods in power system stability studies

In this thesis various mathematical methods of studying the transient and dynamic stabiIity of practical power systems are presented. Certain long established methods are reviewed and refinements of some proposed. New methods are presented which remove some of the difficulties encountered in applying the powerful stability theories based on the concepts of Liapunov. Chapter 1 is concerned with numerical solution of the transient stability problem. Following a review and comparison of synchronous machine models the superiority of a particular model from the point of view of combined computing time and accuracy is demonstrated. A digital computer program incorporating all the synchronous machine models discussed, and an induction machine model, is described and results of a practical multi-machine transient stability study are presented. Chapter 2 reviews certain concepts and theorems due to Liapunov. In Chapter 3 transient stability regions of single, two and multi~machine systems are investigated through the use of energy type Liapunov functions. The treatment removes several mathematical difficulties encountered in earlier applications of the method. In Chapter 4 a simple criterion for the steady state stability of a multi-machine system is developed and compared with established criteria and a state space approach. In Chapters 5, 6 and 7 dynamic stability and small signal dynamic response are studied through a state space representation of the system. In Chapter 5 the state space equations are derived for single machine systems. An example is provided in which the dynamic stability limit curves are plotted for various synchronous machine representations. In Chapter 6 the state space approach is extended to multi~machine systems. To draw conclusions concerning dynamic stability or dynamic response the system eigenvalues must be properly interpreted, and a discussion concerning correct interpretation is included. Chapter 7 presents a discussion of the optimisation of power system small sjgnal performance through the use of Liapunov functions.

Can we observe collective neuronal activity from macroscopic aggregate signals?

The fundamental problem faced by noninvasive neuroimaging techniques such as EEG/MEG1 is to elucidate functionally important aspects of the microscopic neuronal network dynamics from macroscopic aggregate measurements. Due to the mixing of the activities of large neuronal populations in the observed macroscopic aggregate, recovering the underlying network that generates the signal in the absence of any additional information represents a considerable challenge. Recent MEG studies have shown that macroscopic measurements contain sufficient information to allow the differentiation between patterns of activity, which are likely to represent different stimulus-specific collective modes in the underlying network (Hadjipapas, A., Adjamian, P., Swettenham, J.B., Holliday, I.E., Barnes, G.R., 2007. Stimuli of varying spatial scale induce gamma activity with distinct temporal characteristics in human visual cortex. NeuroImage 35, 518–530). The next question arising in this context is whether aspects of collective network activity can be recovered from a macroscopic aggregate signal. We propose that this issue is most appropriately addressed if MEG/EEG signals are to be viewed as macroscopic aggregates arising from networks of coupled systems as opposed to aggregates across a mass of largely independent neural systems. We show that collective modes arising in a network of simulated coupled systems can be indeed recovered from the macroscopic aggregate. Moreover, we show that nonlinear state space methods yield a good approximation of the number of effective degrees of freedom in the network. Importantly, information about hidden variables, which do not directly contribute to the aggregate signal, can also be recovered. Finally, this theoretical framework can be applied to experimental MEG/EEG data in the future, enabling the inference of state dependent changes in the degree of local synchrony in the underlying network.

Self-sustained clusters and ergodicity breaking in spin models

Self-sustained spin clusters are analytically linked to ergodicity breaking in fully connected Ising and Sherrington-Kirkpatick (SK) models, relating the less understood spin space to the well understood state space. This correspondence is established through the absence of clusters in the paramagnetic phase, the presence of one dominant cluster in the Ising ferromagnet, and the formation of nontrivial clusters in SK spin glass. Yet unobserved phenomena are also revealed such as a first order phase transition in cluster sizes in the SK ferromagnet. The method could be adapted to investigate other spin models. © 2013 American Physical Society.

Transition of planar Couette flow at infinite Reynolds numbers

An outline of the state space of planar Couette flow at high Reynolds numbers (Re<105) is investigated via a variety of efficient numerical techniques. It is verified from nonlinear analysis that the lower branch of the hairpin vortex state (HVS) asymptotically approaches the primary (laminar) state with increasing Re. It is also predicted that the lower branch of the HVS at high Re belongs to the stability boundary that initiates a transition to turbulence, and that one of the unstable manifolds of the lower branch of HVS lies on the boundary. These facts suggest HVS may provide a criterion to estimate a minimum perturbation arising transition to turbulent states at the infinite Re limit. © 2013 American Physical Society.

An edge-based matching kernel through discrete-time quantum walks

In this paper, we propose a new edge-based matching kernel for graphs by using discrete-time quantum walks. To this end, we commence by transforming a graph into a directed line graph. The reasons of using the line graph structure are twofold. First, for a graph, its directed line graph is a dual representation and each vertex of the line graph represents a corresponding edge in the original graph. Second, we show that the discrete-time quantum walk can be seen as a walk on the line graph and the state space of the walk is the vertex set of the line graph, i.e., the state space of the walk is the edges of the original graph. As a result, the directed line graph provides an elegant way of developing new edge-based matching kernel based on discrete-time quantum walks. For a pair of graphs, we compute the h-layer depth-based representation for each vertex of their directed line graphs by computing entropic signatures (computed from discrete-time quantum walks on the line graphs) on the family of K-layer expansion subgraphs rooted at the vertex, i.e., we compute the depth-based representations for edges of the original graphs through their directed line graphs. Based on the new representations, we define an edge-based matching method for the pair of graphs by aligning the h-layer depth-based representations computed through the directed line graphs. The new edge-based matching kernel is thus computed by counting the number of matched vertices identified by the matching method on the directed line graphs. Experiments on standard graph datasets demonstrate the effectiveness of our new kernel.

Inference of helicopter airframe condition

The goal of this paper is to model normal airframe conditions for helicopters in order to detect changes. This is done by inferring the flying state using a selection of sensors and frequency bands that are best for discriminating between different states. We used non-linear state-space models (NLSSM) for modelling flight conditions based on short-time frequency analysis of the vibration data and embedded the models in a switching framework to detect transitions between states. We then created a density model (using a Gaussian mixture model) for the NLSSM innovations: this provides a model for normal operation. To validate our approach, we used data with added synthetic abnormalities which was detected as low-probability periods. The model of normality gave good indications of faults during the flight, in the form of low probabilities under the model, with high accuracy (>92 %). © 2013 IEEE.

Fundamental characteristics of space-charge waves in photorefractive sillenite crystals

We investigate experimentally the fundamental characteristics of space-charge waves excited in a photorefractive crystal of Bi12SiO20. Features such as their transient rise and decay as well as their steady-state frequency response are investigated. Based on this, we find the dependence of the space-charge waves' quality factor on spatial frequency and electric-field biasing. The experimental findings are compared with the linear space-charge wave theory developed previously by Sturman et al. [J. Opt. Sec. Am. B 10, 1919 (1993)].