Regularization techniques for ill-posed inverse problems in data assimilation

Optimal state estimation from given observations of a dynamical system by data assimilation is generally an ill-posed inverse problem. In order to solve the problem, a standard Tikhonov, or L2, regularization is used, based on certain statistical assumptions on the errors in the data. The regularization term constrains the estimate of the state to remain close to a prior estimate. In the presence of model error, this approach does not capture the initial state of the system accurately, as the initial state estimate is derived by minimizing the average error between the model predictions and the observations over a time window. Here we examine an alternative L1 regularization technique that has proved valuable in image processing. We show that for examples of flow with sharp fronts and shocks, the L1 regularization technique performs more accurately than standard L2 regularization.

A comparison between classical techniques and eigenstructure assignment in the design of a longitudinal instrument landing system

Although modern control techniques such as eigenstructure assignment have been given extensive coverage in control literature there is a reluctance to use them in practice as they are often not believed to be as `visible' or as simple as classical methods. A simple aircraft example is used, and it is shown that eigenstructure assignment can be used easily to produce a more viable controller than with simple classical techniques.

Virulotyping and antimicrobial resistance typing of salmonella enterica serovars relevant to human health in Europe

The combination of virulence gene and antimicrobial resistance gene typing using DNA arrays is a recently developed genomics-based approach to bacterial molecular epidemiology. We have now applied this technology to 523 Salmonella enterica subsp. enterica strains collected from various host sources and public health and veterinary institutes across nine European countries. The strain set included the five predominant Salmonella serovars isolated in Europe (Enteritidis, Typhimurium, Infantis, Virchow, and Hadar). Initially, these strains were screened for 10 potential virulence factors (avrA, ssaQ, mgtC, siiD, sopB, gipA, sodC1, sopE1, spvC, and bcfC) by polymerase chain reaction. The results indicated that only 14 profiles comprising these genes (virulotypes) were observed throughout Europe. Moreover, most of these virulotypes were restricted to only one (n = 9) or two (n = 4) serovars. The data also indicated that the virulotype did not vary significantly with host source or geographical location. Subsequently, a representative subset of 77 strains was investigated using a microarray designed to detect 102 virulence and 49 resistance determinants. The results confirmed and extended the previous observations using the virulo-polymerase chain reaction screen. Strains belonging to the same serovar grouped together, indicating that the broader virulence-associated gene complement corresponded with the serovar. There were, however, some differences in the virulence gene profiles between strains belonging to an individual serovar. This variation occurred primarily within those virulence genes that were prophage encoded, in fimbrial clusters or in the virulence plasmid. It seems likely that such changes enable Salmonella to adapt to different environmental conditions, which might be reflected in serovar-specific ecology. In this strain subset a number of resistance genes were detected and were serovar restricted to a varying degree. Once again the profiles of those genes encoding resistance were similar or the same for each serovar in all hosts and countries investigated.

Restoring species-rich grassland: principles and techniques

Grassland restoration is the dominant activity funded by agri-environment schemes (AES). However, the re-instatement of biodiversity and ecosystem services is limited by a number of severe abiotic and biotic constraints resulting from previous agricultural management. These appear to be less severe on ex-arable sites compared with permanent grassland. We report findings of a large research programme into practical solutions to these constraints. The key abiotic constraint was high residual soil fertility, particularly phosphorus. This can most easily be addressed by targeting of sites of low nutrient status. The chief biotic constraints were lack of propagules of desirable species and suitable sites for their establishment. Addition of seed mixtures or green hay to gaps created by either mechanical disturbance or herbicide was the most effective means of overcoming these factors. Finally, manipulation of biotic interactions, including hemiparasitic plants to reduce competition from grasses and control of mollusc herbivory of sown species, significantly improved the effectiveness of these techniques.

Bibliographic analysis of persuasive systems: techniques, methods and domains of application

This paper presents findings of our study on peer-reviewed papers published in the International Conference on Persuasive Technology from 2006 to 2010. The study indicated that out of 44 systems reviewed, 23 were reported to be successful, 2 to be unsuccessful and 19 did not specify whether or not it was successful. 56 different techniques were mentioned and it was observed that most designers use ad hoc definitions for techniques or methods used in design. Hence we propose the need for research to establish unambiguous definitions of techniques and methods in the field.

Molecular typing of Salmonella serotypes prevalent in animals in England: assessment of methodology

Salmonella enterica serotypes Derby, Mbandaka, Montevideo, Livingstone, and Senftenberg were among the 10 most prevalent serotypes isolated from farm animals in England and Wales in 1999. These serotypes are of potential zoonotic relevance; however, there is currently no "gold standard" fingerprinting method for them. A collection of isolates representing the former serotypes and serotype Gold Coast were analyzed using plasmid profiling, pulsed-field gel electrophoresis (PFGE), and ribotyping. The success of the molecular methods in identifying DNA polymorphisms was different for each serotype. Plasmid profiling was particularly useful for serotype Derby isolates, and it also provided a good level of discrimination for serotype Senftenberg. For most serotypes, we observed a number of nontypeable plasmid-free strains, which represents a limitation of this technique. Fingerprinting of genomic DNA by ribotyping and PFGE produced a significant variation in results, depending on the serotype of the strain. Both PstI/SphI ribotyping and XbaI-PFGE provided a similar degree of strain differentiation for serotype Derby and serotype Senftenberg, only marginally lower than that achieved by plasmid profiling. Ribotyping was less sensitive than PFGE when applied to serotype Mbandaka or serotype Montevideo. Serotype Gold Coast isolates were found to be nontypeable by XbaI-PFGE, and a significant proportion of them were found to be plasmid free. A similar situation applies to a number of serotype Livingstone isolates which were nontypeable by plasmid profiling and/or PFGE. In summary, the serotype of the isolates has a considerable influence in deciding the best typing strategy; a single method cannot be relied upon for discriminating between strains, and a combination of typing methods allows further discrimination.

Combined use of two genetic fingerprinting methods, pulsed-field gel electrophoresis and ribotyping, for characterization of Escherichia coli O157 isolates from food animals, retail meats, and cases of human disease

Two genetic fingerprinting techniques, pulsed-field gel electrophoresis (PFGE) and ribotyping, were used to characterize 207 Escherichia coli O157 isolates from food animals, foods of animal origin, and cases of human disease (206 of the isolates were from the United Kingdom). In addition, 164 of these isolates were also phage typed. The isolates were divided into two general groups: (i) unrelated isolates not known to be epidemiologically linked (n = 154) and originating from food animals, foods and the environment, or humans and (ii) epidemiologically related isolates (n = 53) comprised of four related groups (RGs) originating either from one farm plus the abattoir where cattle from that farm were slaughtered or from one of three different English abattoirs. PFGE was conducted with the restriction endonuclease XbaI. while for ribotyping, two restriction endonucleases (PstI and SphI) were combined to digest genomic DNAs simultaneously. The 207 E. coli O157 isolates produced 97 PFGE profiles and 51 ribotypes. The two genetic fingerprinting methods had similar powers to discriminate the 154 epidemiologically unrelated E. coli O157 isolates in the study (Simpson's index of diversity [D] = 0.98 and 0.94 for PFGE typing and ribotyping, respectively). There was no correlation between the source of an isolate (healthy meat or milk animals, retail meats, or cases of human infection) and either particular PFGE or ribotype profiles or clusters. Combination of the results of both genetic fingerprinting methods produced 146 types, significantly more than when either of the two methods was used individually. Consequently, the superior discriminatory performance of the PFGE-ribotyping combination was proven in two ways: (i) by demonstrating that the majority of the E. coli O157 isolates with unrelated histories were indeed distinguishable types and (ii) by identifying some clonal groups among two of the four RGs of E. coli O157 isolates (comprising PFGE types different by just one or two bands), the relatedness of which would have remained unconfirmed otherwise.

Scaling up data mining techniques to large datasets using parallel and distributed processing

Advances in hardware and software technology enable us to collect, store and distribute large quantities of data on a very large scale. Automatically discovering and extracting hidden knowledge in the form of patterns from these large data volumes is known as data mining. Data mining technology is not only a part of business intelligence, but is also used in many other application areas such as research, marketing and financial analytics. For example medical scientists can use patterns extracted from historic patient data in order to determine if a new patient is likely to respond positively to a particular treatment or not; marketing analysts can use extracted patterns from customer data for future advertisement campaigns; finance experts have an interest in patterns that forecast the development of certain stock market shares for investment recommendations. However, extracting knowledge in the form of patterns from massive data volumes imposes a number of computational challenges in terms of processing time, memory, bandwidth and power consumption. These challenges have led to the development of parallel and distributed data analysis approaches and the utilisation of Grid and Cloud computing. This chapter gives an overview of parallel and distributed computing approaches and how they can be used to scale up data mining to large datasets.

A comparison of automated keyphrase extraction techniques and of automatic evaluation vs. human evaluation

Keyphrases are added to documents to help identify the areas of interest they contain. However, in a significant proportion of papers author selected keyphrases are not appropriate for the document they accompany: for instance, they can be classificatory rather than explanatory, or they are not updated when the focus of the paper changes. As such, automated methods for improving the use of keyphrases are needed, and various methods have been published. However, each method was evaluated using a different corpus, typically one relevant to the field of study of the method’s authors. This not only makes it difficult to incorporate the useful elements of algorithms in future work, but also makes comparing the results of each method inefficient and ineffective. This paper describes the work undertaken to compare five methods across a common baseline of corpora. The methods chosen were Term Frequency, Inverse Document Frequency, the C-Value, the NC-Value, and a Synonym based approach. These methods were analysed to evaluate performance and quality of results, and to provide a future benchmark. It is shown that Term Frequency and Inverse Document Frequency were the best algorithms, with the Synonym approach following them. Following these findings, a study was undertaken into the value of using human evaluators to judge the outputs. The Synonym method was compared to the original author keyphrases of the Reuters’ News Corpus. The findings show that authors of Reuters’ news articles provide good keyphrases but that more often than not they do not provide any keyphrases.