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.

Sparse representation for Gaussian process models

We develop an approach for a sparse representation for Gaussian Process (GP) models in order to overcome the limitations of GPs caused by large data sets. The method is based on a combination of a Bayesian online algorithm together with a sequential construction of a relevant subsample of the data which fully specifies the prediction of the model. Experimental results on toy examples and large real-world datasets indicate the efficiency of the approach.

Time delay estimation with hidden Markov models

Most traditional methods for extracting the relationships between two time series are based on cross-correlation. In a non-linear non-stationary environment, these techniques are not sufficient. We show in this paper how to use hidden Markov models to identify the lag (or delay) between different variables for such data. Adopting an information-theoretic approach, we develop a procedure for training HMMs to maximise the mutual information (MMI) between delayed time series. The method is used to model the oil drilling process. We show that cross-correlation gives no information and that the MMI approach outperforms maximum likelihood.

Public key cryptography and error correcting codes as Ising models

We employ the methods of statistical physics to study the performance of Gallager type error-correcting codes. In this approach, the transmitted codeword comprises Boolean sums of the original message bits selected by two randomly-constructed sparse matrices. We show that a broad range of these codes potentially saturate Shannon's bound but are limited due to the decoding dynamics used. Other codes show sub-optimal performance but are not restricted by the decoding dynamics. We show how these codes may also be employed as a practical public-key cryptosystem and are of competitive performance to modern cyptographical methods.

Online approximations for wind-field models

We study online approximations to Gaussian process models for spatially distributed systems. We apply our method to the prediction of wind fields over the ocean surface from scatterometer data. Our approach combines a sequential update of a Gaussian approximation to the posterior with a sparse representation that allows to treat problems with a large number of observations.

A novel approach to modelling and exploiting uncertainty in stochastic control systems

We consider an inversion-based neurocontroller for solving control problems of uncertain nonlinear systems. Classical approaches do not use uncertainty information in the neural network models. In this paper we show how we can exploit knowledge of this uncertainty to our advantage by developing a novel robust inverse control method. Simulations on a nonlinear uncertain second order system illustrate the approach.

Improved robust control of nonlinear stochastic systems using uncertain models

We introduce a technique for quantifying and then exploiting uncertainty in nonlinear stochastic control systems. The approach is suboptimal though robust and relies upon the approximation of the forward and inverse plant models by neural networks, which also estimate the intrinsic uncertainty. Sampling from the resulting Gaussian distributions of the inversion based neurocontroller allows us to introduce a control law which is demonstrably more robust than traditional adaptive controllers.

Handling and communicating uncertainty in chained geospatial web services

Recent developments in service-oriented and distributed computing have created exciting opportunities for the integration of models in service chains to create the Model Web. This offers the potential for orchestrating web data and processing services, in complex chains; a flexible approach which exploits the increased access to products and tools, and the scalability offered by the Web. However, the uncertainty inherent in data and models must be quantified and communicated in an interoperable way, in order for its effects to be effectively assessed as errors propagate through complex automated model chains. We describe a proposed set of tools for handling, characterizing and communicating uncertainty in this context, and show how they can be used to 'uncertainty- enable' Web Services in a model chain. An example implementation is presented, which combines environmental and publicly-contributed data to produce estimates of sea-level air pressure, with estimates of uncertainty which incorporate the effects of model approximation as well as the uncertainty inherent in the observational and derived data.

Forecasting exchange rates with linear and nonlinear models

In this paper, the exchange rate forecasting performance of neural network models are evaluated against the random walk, autoregressive moving average and generalised autoregressive conditional heteroskedasticity models. There are no guidelines available that can be used to choose the parameters of neural network models and therefore, the parameters are chosen according to what the researcher considers to be the best. Such an approach, however,implies that the risk of making bad decisions is extremely high, which could explain why in many studies, neural network models do not consistently perform better than their time series counterparts. In this paper, through extensive experimentation, the level of subjectivity in building neural network models is considerably reduced and therefore giving them a better chance of Forecasting exchange rates with linear and nonlinear models 415 performing well. The results show that in general, neural network models perform better than the traditionally used time series models in forecasting exchange rates.

DEA models for ratio data:convexity consideration

Data envelopment analysis (DEA) is defined based on observed units and by finding the distance of each unit to the border of estimated production possibility set (PPS). The convexity is one of the underlying assumptions of the PPS. This paper shows some difficulties of using standard DEA models in the presence of input-ratios and/or output-ratios. The paper defines a new convexity assumption when data includes a ratio variable. Then it proposes a series of modified DEA models which are capable to rectify this problem.

Forecasting exchange rates with linear and nonlinear models

In this paper the exchange rate forecasting performance of neural network models are evaluated against random walk and a range of time series models. There are no guidelines available that can be used to choose the parameters of neural network models and therefore the parameters are chosen according to what the researcher considers to be the best. Such an approach, however, implies that the risk of making bad decisions is extremely high which could explain why in many studies neural network models do not consistently perform better than their time series counterparts. In this paper through extensive experimentation the level of subjectivity in building neural network models is considerably reduced and therefore giving them a better chance of performing well. Our results show that in general neural network models perform better than traditionally used time series models in forecasting exchange rates.

Emulation of dynamic computer models with multivariate output

This preliminary report describes work carried out as part of work package 1.2 of the MUCM research project. The report is split in two parts: the ?rst part (Sections 1 and 2) summarises the state of the art in emulation of computer models, while the second presents some initial work on the emulation of dynamic models. In the ?rst part, we describe the basics of emulation, introduce the notation and put together the key results for the emulation of models with single and multiple outputs, with or without the use of mean function. In the second part, we present preliminary results on the chaotic Lorenz 63 model. We look at emulation of a single time step, and repeated application of the emulator for sequential predic- tion. After some design considerations, the emulator is compared with the exact simulator on a number of runs to assess its performance. Several general issues related to emulating dynamic models are raised and discussed. Current work on the larger Lorenz 96 model (40 variables) is presented in the context of dimension reduction, with results to be provided in a follow-up report. The notation used in this report are summarised in appendix.

Sensitivity of LPGs in PCFs fabricated by an electric arc to temperature, strain, and external refractive index

Sensing properties of long-period gratings (LPGs) fabricated in photonic crystal fibers by an electric arc are explained and quantified by semianalytical and numerical models. In particular, the grating's insensitivity to temperature and simultaneous sensitivity to strain and refractive index are simulated. The modeling procedure is generalized so that it can be applied to a wide range of LPGs in various fibers.

Employee negative emotions in supervisory relationships : an attachment theory perspective

Relationships with supervisors are a major source of negative emotions at work, but little is known about why this is so. The aim of the research was to use attachment theory (Bowlby, 1969, 1973; 1980) as a framework for investigating the nature and causes of employee negative emotional experiences, in the context of their supervisory relationships. The research was conducted in three stages. In Stage 1 two studies were conducted to develop a measure of employee perceptions of supervisor caregiving (SCS). Results indicated that the 20-item scale had good reliability and validity. Stage 2 required participants (N=183) to complete a questionnaire that was designed to examine the roles of supervisor caregiving and working models (specific and global) in determining cognitive and emotional responses to hypothetical supervisor behaviours. The results provided partial support for an Independent Effects Model. Supervisor caregiving predicted specific anxiety and avoidance. In tum, both dimensions of attachment predicted negative emotions, but this relationship was mediated by event interpretation only in the case of avoidance. Global models made a smaller but significant contribution to negative emotions overall. There was no support for an interaction effect between specific and global models in determining event interpretation. In stage 3 a sub-sample of questionnaire respondents (N=24) were interviewed about 'real-life' caregiving and negative emotional experiences in their supervisory relationships. Secure individuals experienced supervisors as consistently warm, available, and responsive. They reported few negative events or emotions. Individuals with insecure specific working models experienced rejecting or inconsistent supervisor caregiving. They were sensitised to trust and closeness issues in their relationships, and reported negative events and emotions underpinned by these themes. Overall, results broadly supported attachment theory predictions. It is concluded that an attachment theory perspective provides new insight into the nature and causes of employee negative emotions in supervisory relationships.

Trend analysis in scientometrics

The thesis investigates the properties of two trends or time series which formed a:part of the Co-Citation bibliometric model "X~Ray Crystallography and Protein Determination in 1978, 1980 and 1982". This model was one of several created for the 1983 ABRC Science Policy Study which aimed to test the utility of bibliometric models in a national science policy context. The outcome of the validation part of that study proved to be especially favourable concerning the utility of trend data, which purport to model the development of speciality areas in science over time. This assessment could have important implications for the use of such data in policy formulation. However one possible problem with the Science Policy Study's conclusions was that insufficient time was available in the study for an in-depth analysis of the data. The thesis aims to continue the validation begun in the ABRC study by providing a detailed.examination of the characteristics of the data contained in the Trends numbered 11 and 44 in the model. A novel methodology for the analysis of the properties of the trends with respect to their literature content is presented. This is followed by an assessment based on questionnaire and interview data, of the ability of Trend 44 to realistically model the historical development of the field of mobile genetic elements research over time, with respect to its scientific content and the activities of its community of researchers. The results of these various analyses are then used to evaluate the strenghts and weaknesses of a trend or time series approach to the modelling of the activities of scientifiic fields. A critical evaluation of the origins of the discovered strengths and weaknesses.in the assumptions underlying the techniques used to generate trends from co-citation data is provided. Possible improvements. to the modelling techniques are discussed.