Efficient accommodation of local minima in watershed model calibration

The Gauss-Marquardt-Levenberg (GML) method of computer-based parameter estimation, in common with other gradient-based approaches, suffers from the drawback that it may become trapped in local objective function minima, and thus report optimized parameter values that are not, in fact, optimized at all. This can seriously degrade its utility in the calibration of watershed models where local optima abound. Nevertheless, the method also has advantages, chief among these being its model-run efficiency, and its ability to report useful information on parameter sensitivities and covariances as a by-product of its use. It is also easily adapted to maintain this efficiency in the face of potential numerical problems (that adversely affect all parameter estimation methodologies) caused by parameter insensitivity and/or parameter correlation. The present paper presents two algorithmic enhancements to the GML method that retain its strengths, but which overcome its weaknesses in the face of local optima. Using the first of these methods an intelligent search for better parameter sets is conducted in parameter subspaces of decreasing dimensionality when progress of the parameter estimation process is slowed either by numerical instability incurred through problem ill-posedness, or when a local objective function minimum is encountered. The second methodology minimizes the chance of successive GML parameter estimation runs finding the same objective function minimum by starting successive runs at points that are maximally removed from previous parameter trajectories. As well as enhancing the ability of a GML-based method to find the global objective function minimum, the latter technique can also be used to find the locations of many non-global optima (should they exist) in parameter space. This can provide a useful means of inquiring into the well-posedness of a parameter estimation problem, and for detecting the presence of bimodal parameter and predictive probability distributions. The new methodologies are demonstrated by calibrating a Hydrological Simulation Program-FORTRAN (HSPF) model against a time series of daily flows. Comparison with the SCE-UA method in this calibration context demonstrates a high level of comparative model run efficiency for the new method. (c) 2006 Elsevier B.V. All rights reserved.

The effect of St John's wort extracts on CYP3A: a systematic review of prospective clinical trials

Aim The aim of this systematic review was to assess the quality and outcomes of clinical trials investigating the effect of St John's wort extracts on the metabolism of drugs by CYP3A. Methods Prospective clinical trials assessing the effect of St John's wort (SJW) extracts on metabolism by CYP3A were identified through computer-based searches (from their inception to May 2005) of Medline, Cinahl, PsycINFO, AMED, Current Contents and Embase, hand-searches of bibliographies of relevant papers and consultation with manufacturers and researchers in the field. Two reviewers selected trials for inclusion, independently extracted data and recorded details on study design. Results Thirty-one studies met the eligibility criteria. More than two-thirds of the studies employed a before-and-after design, less than one-third of the studies used a crossover design, and only three studies were double-blind and placebo controlled. In 12 studies the SJW extract had been assayed, and 14 studies stated the specific SJW extract used. Results from 26 studies, including all of the 19 studies that used high-dose hyperforin extracts (> 10 mg day(-1)), had outcomes consistent with CYP3A induction. The three studies using low-dose hyperforin extracts (< 4 mg day(-1)) demonstrated no significant effect on CYP3A. Conclusion There is reasonable evidence to suggest that high-dose hyperforin SJW extracts induce CYP3A. More studies are required to determine whether decreased CYP3A induction occurs after low-dose hyperforin extracts. Future studies should adopt study designs with a control phase or control group, identify the specific SJW extract employed and provide quantitative analyses of key constituents.

The role of social capital in the promotion of conservation farming: The case of 'landcare' in the southern Philippines

'Social capital' refers to the relationships of trust, communication, and cooperation that facilitate collective action in a community. It is particularly relevant to soil conservation in developing countries, which requires collective efforts to raise awareness of soil degradation, provide effective training in soil conservation practices, and implement soil conservation measures on individual farms. The Landcare Program in the Southern Philippines promotes simple conservation practices in upland environments through establishing and supporting community landcare groups and municipal landcare associations, thus augmenting the social capital of farmers in these locations. An evaluation of the Landcare Program in Barangay Ned, South Cotabato, based on a survey of 313 farm households and case studies of nine landcare groups, shows that, despite extreme isolation and difficult working conditions, farmers responded by rapidly forming landcare groups and a landcare association, and adopting contour barriers on their maize farms. They utilized the bonding social capital inhering in their local communities to build stocks of bridging social capital, linking them to information, training and resources from outside their immediate locality. A logistic regression model of the factors affecting adoption of contour barriers shows that farmers who had undergone the practical, farmer-based training provided by the Landcare Program, and who were members of a landcare group, were significantly more likely to adopt conservation measures. These results confirm the value of investing in social capital to promote soil conservation. Copyright (c) 2005 John Wiley & Sons, Ltd.

Automatic measurement of vertebral rotation in idiopathic scoliosis

Study Design. Development of an automatic measurement algorithm and comparison with manual measurement methods. Objectives. To develop a new computer-based method for automatic measurement of vertebral rotation in idiopathic scoliosis from computed tomography images and to compare the automatic method with two manual measurement techniques. Summary of Background Data. Techniques have been developed for vertebral rotation measurement in idiopathic scoliosis using plain radiographs, computed tomography, or magnetic resonance images. All of these techniques require manual selection of landmark points and are therefore subject to interobserver and intraobserver error. Methods. We developed a new method for automatic measurement of vertebral rotation in idiopathic scoliosis using a symmetry ratio algorithm. The automatic method provided values comparable with Aaro and Ho's manual measurement methods for a set of 19 transverse computed tomography slices through apical vertebrae, and with Aaro's method for a set of 204 reformatted computed tomography images through vertebral endplates. Results. Confidence intervals (95%) for intraobserver and interobserver variability using manual methods were in the range 5.5 to 7.2. The mean (+/- SD) difference between automatic and manual rotation measurements for the 19 apical images was -0.5 degrees +/- 3.3 degrees for Aaro's method and 0.7 degrees +/- 3.4 degrees for Ho's method. The mean (+/- SD) difference between automatic and manual rotation measurements for the 204 endplate images was 0.25 degrees +/- 3.8 degrees. Conclusions. The symmetry ratio algorithm allows automatic measurement of vertebral rotation in idiopathic scoliosis without intraobserver or interobserver error due to landmark point selection.

Employment services in an age of e-government

The increasing use of information and communications technologies among government departments and non-government agencies has fundamentally changed the implementation of employment services policy in Australia. The administrative arrangements for governing unemployment and unemployed people are now constituted by a complex contractual interplay between government departments as ‘purchasers’ and a range of small and large private organizations as ‘providers’. Assessing, tracking and monitoring the activities of unemployed people through the various parts of the employment services system has been made possible by developments in information technology and tailored computer programs. Consequently, the discretionary capacity that is traditionally associated with ‘street-level bureaucracy’ has been partly transformed into more prescriptive forms of ‘screen-level bureaucracy’. The knowledge embedded in these new computer-based technologies is considered superior because it is based on ‘objective calculations’, rather than subjective assessments of individual employees. The relationship between the sociopolitical context of unemployment policy and emerging forms of e-government is explored using illustrative findings from a qualitative pilot study undertaken in two Australian sites. The findings suggest that some of the new technologies in the employment services system are welcomed, while other applications are experienced as contradictory to the aims of delivering a personalized and respectful service.

Interference between propagating changes and its effect on system failure

Computer-based, socio-technical systems projects are frequently failures. In particular, computer-based information systems often fail to live up to their promise. Part of the problem lies in the uncertainty of the effect of combining the subsystems that comprise the complete system; i.e. the system's emergent behaviour cannot be predicted from a knowledge of the subsystems. This paper suggests uncertainty management is a fundamental unifying concept in analysis and design of complex systems and goes on to indicate that this is due to the co-evolutionary nature of the requirements and implementation of socio-technical systems. The paper shows a model of the propagation of a system change that indicates that the introduction of two or more changes over time can cause chaotic emergent behaviour.

Anthropometric-based selection and sprint kayak training in children

A 12 week kayak training programme was evaluated in children who either had or did not have the anthropometric characteristics identified as being unique to senior elite sprint kayakers. Altogether, 234 male and female school children were screened to select 10 children with and 10 children without the identified key anthropometric characteristics. Before and after training, the children completed an all-out 2 min kayak ergometer simulation test; measures of oxygen consumption, plasma lactate and total work accomplished were recorded. In addition, a 500 m time trial was performed at weeks 3 and 12. The coaches were unaware which 20 children possessed those anthropometric characteristics deemed to favour development of kayak ability. All children improved in both the 2 min ergometer simulation test and 500 m time trial. However, boys who were selected according to favourable anthropometric characteristics showed greater improvement than those without such characteristics in the 2 min ergometer test only. In summary, in a small group of children selected according to anthropometric data unique to elite adult kayakers, 12 weeks of intensive kayak training did not influence the rate of improvement of on-water sprint kayak performance.

Can anticipatory skills be learned through implicit video-based perceptual training?

The aim of this experiment was to determine the effectiveness of two video-based perceptual training approaches designed to improve the anticipatory skills of junior tennis players. Players were assigned equally to an explicit learning group, an implicit learning group, a placebo group or a control group. A progressive temporal occlusion paradigm was used to examine, before and after training, the ability of the players to predict the direction of an opponent's service in an in-vivo on-court setting. The players responded either through hitting a return stroke or making a verbal prediction of stroke direction. Results revealed that the implicit learning group, whose training required them to predict serve speed direction while viewing temporally occluded video footage of the return-of-serve scenario, significantly improved their prediction accuracy after the training intervention. However, this training effect dissipated after a 32 day unfilled retention interval. The explicit learning group, who received instructions about the specific aspects of the pre-contact service kinematics that are informative with respect to service direction, did not demonstrate any significant performance improvements after the intervention. This, together with the absence of any significant improvements for the placebo and control groups, demonstrated that the improvement observed for the implicit learning group was not a consequence of either expectancy or familiarity effects.

Parent management training and Asperger syndrome - A randomized controlled trial to evaluate a parent based intervention

This controlled trial of a parent management intervention aimed to increase parental competence in management of problem behaviours associated with Asperger syndrome. The intervention compared two formats, a 1 day workshop and six individual sessions. Measures were taken on three occasions: pre-intervention, at 4 weeks, and at 3 month follow-up. Variables of interest were number and intensity of problem behaviours, and parent evaluation of social interaction skills. Results showed parents reporting fewer and lower intensity of problem behaviours and increased social interactions at 4 weeks and 3 months. Results held across formats and suggest that parent management training can provide an effective intervention for parents of a child with Asperger syndrome. Group differences on outcome measures and in the use of strategies are discussed along with limitations of the study.

Training school personnel to implement a universal school-based prevention of depression program under real-world conditions

The present study evaluated the impact of a universal prevention of depression program [the Resourceful Adolescent Program (RAP)] when implemented under real-world conditions in a school setting. Prior research has found the RAP program to be beneficial for high-school students when the program was implemented by university staff selected, trained, and supervised by a research team. The present study evaluated the RAP program when implemented by existing school personnel. Separately, we measured the impact of a training program for facilitators, the quality of subsequent program implementation, and the student's response to the RAP Program. Results showed that, in response to the training program, facilitators believed they had acquired the knowledge and confidence to implement the program and that the quality of program implementation was acceptable. The study did not demonstrate a beneficial impact of the RAP program for the students. The results raise important questions regarding the extent of training and ongoing supervision facilitators require if the beneficial outcomes for students are to be maintained when interventions are implemented under real-world conditions in school settings. (C) 2004 Society for the Study of School Psychology. Published by Elsevier Ltd. All rights reserved.

Theoretical frameworks for analysing turn-based computer strategy games

This paper argues for a more specific formal methodology for the textual analysis of individual game genres. In doing so, it advances a set of formal analytical tools and a theoretical framework for the analysis of turn-based computer strategy games. The analytical tools extend the useful work of Steven Poole, who suggests a Peircian semiotic approach to the study of games as formal systems. The theoretical framework draws upon postmodern cultural theory to analyse and explain the representation of space and the organisation of knowledge in these games. The methodology and theoretical framework is supported by a textual analysis of Civilization II, a significant and influential turn-based computer strategy game. Finally, this paper suggests possibilities for future extensions of this work.

An Incremental EM-based Learning Approach for On-Line Prediction of Hospital Resource Utilization

Objective: Inpatient length of stay (LOS) is an important measure of hospital activity, health care resource consumption, and patient acuity. This research work aims at developing an incremental expectation maximization (EM) based learning approach on mixture of experts (ME) system for on-line prediction of LOS. The use of a batchmode learning process in most existing artificial neural networks to predict LOS is unrealistic, as the data become available over time and their pattern change dynamically. In contrast, an on-line process is capable of providing an output whenever a new datum becomes available. This on-the-spot information is therefore more useful and practical for making decisions, especially when one deals with a tremendous amount of data. Methods and material: The proposed approach is illustrated using a real example of gastroenteritis LOS data. The data set was extracted from a retrospective cohort study on all infants born in 1995-1997 and their subsequent admissions for gastroenteritis. The total number of admissions in this data set was n = 692. Linked hospitalization records of the cohort were retrieved retrospectively to derive the outcome measure, patient demographics, and associated co-morbidities information. A comparative study of the incremental learning and the batch-mode learning algorithms is considered. The performances of the learning algorithms are compared based on the mean absolute difference (MAD) between the predictions and the actual LOS, and the proportion of predictions with MAD < 1 day (Prop(MAD < 1)). The significance of the comparison is assessed through a regression analysis. Results: The incremental learning algorithm provides better on-line prediction of LOS when the system has gained sufficient training from more examples (MAD = 1.77 days and Prop(MAD < 1) = 54.3%), compared to that using the batch-mode learning. The regression analysis indicates a significant decrease of MAD (p-value = 0.063) and a significant (p-value = 0.044) increase of Prop(MAD