Support vector machine classification for large data sets via minimum enclosing ball clustering

Support vector machine (SVM) is a powerful technique for data classification. Despite of its good theoretic foundations and high classification accuracy, normal SVM is not suitable for classification of large data sets, because the training complexity of SVM is highly dependent on the size of data set. This paper presents a novel SVM classification approach for large data sets by using minimum enclosing ball clustering. After the training data are partitioned by the proposed clustering method, the centers of the clusters are used for the first time SVM classification. Then we use the clusters whose centers are support vectors or those clusters which have different classes to perform the second time SVM classification. In this stage most data are removed. Several experimental results show that the approach proposed in this paper has good classification accuracy compared with classic SVM while the training is significantly faster than several other SVM classifiers.

Minimum Wage Regulation under Devolution in Northern Ireland

This paper assesses the development and functioning of regional minimum wage regulation in Northern Ireland in the interwar period under a federal form of devolution. Unlike current devolution arrangements in Scotland and Wales, this gave the Stormont Parliament powers over employment and minimum wage regulation. Northern Ireland Trade Boards were set up by the Ulster Unionist Government under the Trade Boards (Northern Ireland) Act 1923 and functioned along the same lines as those in Great Britain. Uniquely in the UK in this period, employer opposition resulted in the main Trade Board in the Irish Linen Industry being replaced by voluntary collective bargaining machinery. About one-quarter of employees were covered by minimum wage regulation, including two-thirds of females in Belfast, keeping a protective floor under low pay.

First record of minimum Irish hare (Lepus timidus hibernicus Bell 1837) longevity in the wild.

Average longevity of the mountain hare (Lepus timidus L., 1758) has been estimated at nine years in the wild (Macdonald D. and Barrett, P. 1993 Mammals of Britain and Europe. Harper Collins Publishers, London) with a maximum recorded age of 18 years for one marked animal (Angerbjörn, A. and Flux, J. E. C. 1995 Lepus timidus. Mammalian Species 495: 1–11). However, the longevity of the Irish hare (L. t. hibernicus Bell 1837) is entirely unknown. A total of 14 Irish hares was trapped and tagged at Belfast International Airport, Co. Antrim from February to April 2005. The sex, age (juvenile or adult) and weight of each animal were recorded. Adults were taken as those individuals >8-10 months old defined by the fusing of the notch between the apophysis and diaphysis of the tibia and humerus (Flux, J. E. C. 1970 Journal of Zoology 161: 75-123). Individual identification was made by a system of colourcoded ear tags (Roxan iD Ltd. Selkirk, Scotland) being inserted in the centre of the pinna of each ear. Each ear tag (6 × 34 mm) and puncture site was disinfected with 70 per cent ethanol prior to insertion. An adult male, #001/002 ‘Blue/Blue’, was tagged on 3 March 2005 weighing 3.8 kg and was sighted during a return site visit on 4 April 2007. An adult female, #026/003 ‘Green/Yellow’, was tagged on 15 April 2005 weighing 4.0 kg and was sighted during return visits on 25 March 2010 and 19 October 2010. The latest possible date of birth for both individuals was spring/summer 2004. Consequently, they were at least 3 years and 6.5 years old, respectively. This is the first record of minimum Irish hare longevity in the wild. These observations suggest that ear tagging does not compromise animal welfare and is an effective means of long-term monitoring. Future research may utilize capture-mark-recapture methods.

Do minimum wages raise employment? Evidence from the US retail-trade sector

This paper examines the impact of minimum wages on earnings and employment in selected branches of the retail-trade sector, 1990-2005, using county-level data on employment and a panel regression framework that allows for county-specific trends in sectoral outcomes. We focus on specific subsectors within retail trade that are identified as particularly low-wage. We find little evidence of disemployment effects once we allow for geographic-specific trends. Indeed, in many sectors the evidence points to modest (but robust) positive employment effects. (C) 2009 Elsevier B.V. All rights reserved.

Establishing a minimum dataset for prospective registration of systematic reviews: an international consultation

Background: In response to growing recognition of the value of prospective registration of systematic review protocols, we planned to develop a web-based open access international register. In order for the register to fulfil its aims of reducing unplanned duplication, reducing publication bias, and providing greater transparency, it was important to ensure the appropriate data were collected. We therefore undertook a consultation process with experts in the field to identify a minimum dataset for registration. Methods and Findings: A two-round electronic modified Delphi survey design was used. The international panel surveyed included experts from areas relevant to systematic review including commissioners, clinical and academic researchers, methodologists, statisticians, information specialists, journal editors and users of systematic reviews. Direct invitations to participate were sent out to 315 people in the first round and 322 in the second round. Responses to an open invitation to participate were collected separately. There were 194 (143 invited and 51 open) respondents with a 100% completion rate in the first round and 209 (169 invited and 40 open) respondents with a 91% completion rate in the second round. In the second round, 113 (54%) of the participants reported having previously taken part in the first round. Participants were asked to indicate whether a series of potential items should be designated as optional or required registration items, or should not be included in the register. After the second round, a 70% or greater agreement was reached on the designation of 30 of 36 items. Conclusions: The results of the Delphi exercise have established a dataset of 22 required items for the prospective registration of systematic reviews, and 18 optional items. The dataset captures the key attributes of review design as well as the administrative details necessary for registration.

Determining the minimum clinically important differences for outcomes in the DOMINO trial

Although less likely to be reported in clinical trials than expressions of the statistical significance of differences in outcomes, whether or not a treatment has delivered a specified minimum clinically important difference (MCID) is also relevant to patients and their caregivers and doctors. Many dementia treatment randomised controlled trials (RCTs) have not reported MCIDs and, where they have been done, observed differences have not reached these.