Identifying problematic classes in text classification

Real-world text classification tasks often suffer from poor class structure with many overlapping classes and blurred boundaries. Training data pooled from multiple sources tend to be inconsistent and contain erroneous labelling, leading to poor performance of standard text classifiers. The classification of health service products to specialized procurement classes is used to examine and quantify the extent of these problems. A novel method is presented to analyze the labelled data by selectively merging classes where there is not enough information for the classifier to distinguish them. Initial results show the method can identify the most problematic classes, which can be used either as a focus to improve the training data or to merge classes to increase confidence in the predicted results of the classifier.

A novel shape feature for fast region-based pedestrian recognition

A new class of shape features for region classification and high-level recognition is introduced. The novel Randomised Region Ray (RRR) features can be used to train binary decision trees for object category classification using an abstract representation of the scene. In particular we address the problem of human detection using an over segmented input image. We therefore do not rely on pixel values for training, instead we design and train specialised classifiers on the sparse set of semantic regions which compose the image. Thanks to the abstract nature of the input, the trained classifier has the potential to be fast and applicable to extreme imagery conditions. We demonstrate and evaluate its performance in people detection using a pedestrian dataset.

Self-organising neural networks for adaptive control

Self-organizing neural networks have been implemented in a wide range of application areas such as speech processing, image processing, optimization and robotics. Recent variations to the basic model proposed by the authors enable it to order state space using a subset of the input vector and to apply a local adaptation procedure that does not rely on a predefined test duration limit. Both these variations have been incorporated into a new feature map architecture that forms an integral part of an Hybrid Learning System (HLS) based on a genetic-based classifier system. Problems are represented within HLS as objects characterized by environmental features. Objects controlled by the system have preset targets set against a subset of their features. The system's objective is to achieve these targets by evolving a behavioural repertoire that efficiently explores and exploits the problem environment. Feature maps encode two types of knowledge within HLS — long-term memory traces of useful regularities within the environment and the classifier performance data calibrated against an object's feature states and targets. Self-organization of these networks constitutes non-genetic-based (experience-driven) learning within HLS. This paper presents a description of the HLS architecture and an analysis of the modified feature map implementing associative memory. Initial results are presented that demonstrate the behaviour of the system on a simple control task.

Using self-organising feature maps for the control of artificial organisms

Variations on the standard Kohonen feature map can enable an ordering of the map state space by using only a limited subset of the complete input vector. Also it is possible to employ merely a local adaptation procedure to order the map, rather than having to rely on global variables and objectives. Such variations have been included as part of a hybrid learning system (HLS) which has arisen out of a genetic-based classifier system. In the paper a description of the modified feature map is given, which constitutes the HLSs long term memory, and results in the control of a simple maze running task are presented, thereby demonstrating the value of goal related feedback within the overall network.

Adaptive alarm processor for fault diagnosis on power transmission networks

The authors describe a learning classifier system (LCS) which employs genetic algorithms (GA) for adaptive online diagnosis of power transmission network faults. The system monitors switchgear indications produced by a transmission network, reporting fault diagnoses on any patterns indicative of faulted components. The system evaluates the accuracy of diagnoses via a fault simulator developed by National Grid Co. and adapts to reflect the current network topology by use of genetic algorithms.

Evaluation of the time resolved fluorescence immunoassay (TRFIA) for the detection of varicella zoster virus (VZV) antibodies following vaccination of healthcare workers

Determination of varicella zoster virus (VZV) immunity in healthcare workers without a history of chickenpox is important for identifying those in need of vOka vaccination. Post immunisation, healthcare workers in the UK who work with high risk patients are tested for seroconversion. To assess the performance of the time-resolved fluorescence immunoassay (TRFIA) for the detection of antibody in vaccinated as well as unvaccinated individuals, a cut-off was first calculated. VZV-IgG specific avidity and titres six weeks after the first dose of vaccine were used to identify subjects with pre-existing immunity among a cohort of 110 healthcare workers. Those with high avidity (≥60%) were considered to have previous immunity to VZV and those with low or equivocal avidity (<60%) were considered naive. The former had antibody levels ≥400mIU/mL and latter had levels <400mIU/mL. Comparison of the baseline values of the naive and immune groups allowed the estimation of a TRFIA cut-off value of >130mIU/mL which best discriminated between the two groups and this was confirmed by ROC analysis. Using this value, the sensitivity and specificity of TRFIA cut-off were 90% (95% CI 79-96), and 78% (95% CI 61-90) respectively in this population. A subset of samples tested by the gold standard Fluorescence Antibody to Membrane Antigen (FAMA) test showed 84% (54/64) agreement with TRFIA.

A concept of approximated densities for efficient nonlinear estimation

This paper presents the theoretical development of a nonlinear adaptive filter based on a concept of filtering by approximated densities (FAD). The most common procedures for nonlinear estimation apply the extended Kalman filter. As opposed to conventional techniques, the proposed recursive algorithm does not require any linearisation. The prediction uses a maximum entropy principle subject to constraints. Thus, the densities created are of an exponential type and depend on a finite number of parameters. The filtering yields recursive equations involving these parameters. The update applies the Bayes theorem. Through simulation on a generic exponential model, the proposed nonlinear filter is implemented and the results prove to be superior to that of the extended Kalman filter and a class of nonlinear filters based on partitioning algorithms.

Identifying likely successful users of virtual reality systems

This paper suggests a method for identifying individuals who are most suited to using virtual reality (VR) systems. The aim is to help both an individual or employer to decide where that individual's skills and abilities would be best deployed. By considering a potential user's competence and temperament, a graphical representation is introduced that may then be used to crudely delineate a high-aptitude participant against those with lesser capabilities. By introducing standard tests for competence and a standard classifier for temperament, and by further weighting each measure with respect to the technology currently available and the application, a detailed representation of the effectiveness of different users is developed.

Rational expectations? Developer behaviour and development cycles in the central London office market

Purpose – Expectations of future market conditions are acknowledged to be crucial for the development decision and hence for shaping the built environment. The purpose of this paper is to study the central London office market from 1987 to 2009 and test for evidence of rational, adaptive and naive expectations. Design/methodology/approach – Two parallel approaches are applied to test for either rational or adaptive/naive expectations: vector auto-regressive (VAR) approach with Granger causality tests and recursive OLS regression with one-step forecasts. Findings – Applying VAR models and a recursive OLS regression with one-step forecasts, the authors do not find evidence of adaptive and naïve expectations of developers. Although the magnitude of the errors and the length of time lags between market signal and construction starts vary over time and development cycles, the results confirm that developer decisions are explained, to a large extent, by contemporaneous and historic conditions in both the City and the West End, but this is more likely to stem from the lengthy design, financing and planning permission processes rather than adaptive or naive expectations. Research limitations/implications – More generally, the results of this study suggest that real estate cycles are largely generated endogenously rather than being the result of large demand shocks and/or irrational behaviour. Practical implications – Developers may be able to generate excess profits by exploiting market inefficiencies but this may be hindered in practice by the long periods necessary for planning and construction of the asset. Originality/value – This paper focuses the scholarly debate of real estate cycles on the role of expectations. It is also one of very few spatially disaggregate studies of the subject matter.

"Oft expectation fails": a time-series analysis of construction starts in the London office market

Expectations of future market conditions are generally acknowledged to be crucial for the development decision and hence for shaping the built environment. This empirical study of the Central London office market from 1987 to 2009 tests for evidence of adaptive and naive expectations. Applying VAR models and a recursive OLS regression with one-step forecasts, we find evidence of adaptive and naïve, rather than rational expectations of developers. Although the magnitude of the errors and the length of time lags vary over time and development cycles, the results confirm that developers’ decisions are explained to a large extent by contemporaneous and past conditions in both London submarkets. The corollary of this finding is that developers may be able to generate excess profits by exploiting market inefficiencies but this may be hindered in practice by the long periods necessary for planning and construction of the asset. More generally, the results of this study suggest that real estate cycles are largely generated endogenously rather than being the result of unexpected exogenous shocks.

Predicting the annual performance of UK real estate: timeliness versus coverage

The IPD Annual Index is the largest and most comprehensive Real Estate market index available in the UK Such coverage however inevitably leads to delays in publication. In contrast there are a number of quarterly and monthly indices which are published within days of the year end but which lack the coverage in terms of size and numbers of properties. This paper analyses these smaller but more timely indices to see whether such indices can be used to predict the performance of the IPD Annual Index. Using a number of measures of forecasting accuracy it is shown that the smaller indices provide unbiased and efficient predictions of the IPD Annual Index. Such indices also significantly outperform a naive no-change model. Although no one index performs significantly better than the others. The more timely indices however do not perfectly track the IPD Annual Index. As a result any short run predictions of performance will be subject to a degree of error. Nevertheless the more timely indices, although lacking authoritative coverage, provide a valuable service to investors giving good estimates of Real Estates performance well before the publication of the IPD Annual Index.

Prediction and functional analysis of native disorder in proteins from the Three Kingdoms of Life

An automatic method for recognizing natively disordered regions from amino acid sequence is described and benchmarked against predictors that were assessed at the latest critical assessment of techniques for protein structure prediction (CASP) experiment. The method attains a Wilcoxon score of 90.0, which represents a statistically significant improvement on the methods evaluated on the same targets at CASP. The classifier, DISOPRED2, was used to estimate the frequency of native disorder in several representative genomes from the three kingdoms of life. Putative, long (>30 residue) disordered segments are found to occur in 2.0% of archaean, 4.2% of eubacterial and 33.0% of eukaryotic proteins. The function of proteins with long predicted regions of disorder was investigated using the gene ontology annotations supplied with the Saccharomyces genome database. The analysis of the yeast proteome suggests that proteins containing disorder are often located in the cell nucleus and are involved in the regulation of transcription and cell signalling. The results also indicate that native disorder is associated with the molecular functions of kinase activity and nucleic acid binding.

Secondary structure prediction with support vector machines

Motivation: A new method that uses support vector machines (SVMs) to predict protein secondary structure is described and evaluated. The study is designed to develop a reliable prediction method using an alternative technique and to investigate the applicability of SVMs to this type of bioinformatics problem. Methods: Binary SVMs are trained to discriminate between two structural classes. The binary classifiers are combined in several ways to predict multi-class secondary structure. Results: The average three-state prediction accuracy per protein (Q3) is estimated by cross-validation to be 77.07 ± 0.26% with a segment overlap (Sov) score of 73.32 ± 0.39%. The SVM performs similarly to the 'state-of-the-art' PSIPRED prediction method on a non-homologous test set of 121 proteins despite being trained on substantially fewer examples. A simple consensus of the SVM, PSIPRED and PROFsec achieves significantly higher prediction accuracy than the individual methods. Availability: The SVM classifier is available from the authors. Work is in progress to make the method available on-line and to integrate the SVM predictions into the PSIPRED server.

Non-existence of competitive equilibria with dynamically inconsistent preferences

This paper shows the robust non-existence of competitive equilibria even in a simple three period representative agent economy with dynamically inconsistent preferences. We distinguish between a sophisticated and naive representative agent. Even when underlying preferences are monotone and convex, at given prices, we show by example that the induced preference of the sophisticated representative agent over choices in first-period markets is both non-convex and satiated. Even allowing for negative prices, the market-clearing allocation is not contained in the convex hull of demand. Finally, with a naive representative agent, we show that perfect foresight is incompatible with market clearing and individual optimization at given prices.

Fitness and dissemination of disinfectant-selected multiple-antibiotic-resistant (MAR) strains of Salmonella enterica serovar Typhimurium in chickens

Objectives: The aims of this study were to determine whether strains of Salmonella enterica serovar Typhimurium which had acquired low-level multiple antibiotic resistance (MAR) through repeated exposure to farm disinfectants were able to colonize and transmit between chicks as easily as the parent strain and, if such strains were less susceptible to fluoroquinolones, would high-level resistance be selected after fluoroquinolone treatment. Methods: Two mutants were compared with the isogenic parent. In the first experiment, day-old chicks were co-infected with both the parent and a mutant to determine their relative fitness. In the second experiment, parent and mutant strains (in separate groups of chicks) were assessed for their ability to transmit from infected (contact) to non-infected (naive) birds and with respect to their susceptibility to fluoroquinolone treatment. Birds were regularly monitored for the presence of Salmonella in caecal contents. Replica plating was used to monitor for the selection of antibiotic-resistant strains. Results: The parent strain was shown to be significantly fitter than the two mutants and was more rapidly disseminated to naive birds. Antibiotic treatment did not preferentially select for the two mutants or for resistant strains. Conclusions: The disinfectant-exposed strains, although MAR, were less fit, less able to disseminate than the parent strain and were not preferentially selected by therapeutic antibiotic treatment. As such, these strains are unlikely to present a greater problem than other salmonellae in chickens.