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.

Energetic costs of parasitism in the Cape ground squirrel Xerus inauris

Parasites have been suggested to influence many aspects of host behaviour. Some of these effects may be mediated via their impact on host energy budgets. This impact may include effects on both energy intake and absorption as well as components of expenditure, including resting metabolic rate (RMR) and activity (e.g. grooming). Despite their potential importance, the energy costs of parasitism have seldom been directly quantified in a field setting. Here we pharmacologically treated female Cape ground squirrels (Xerus inauris) with anti-parasite drugs and measured the change in body composition, the daily energy expenditure (DEE) using doubly labelled water, the RMR by respirometry and the proportions of time spent looking for food, feeding, moving and grooming. Post-treatment animals gained an average 19 g of fat or approximately 25 kJ d(-1). DEE averaged 382 kJ d-1 prior to and 375 kJ d-1 post treatment (p> 0.05). RMR averaged 174 kJ d-1 prior to and 217 kJ d-1 post treatment (p

Differential energy costs of winter acclimatized common spiny mice Acomys cahirinus from two adjacent habitats

The common spiny mouse Acomys cahirinus, of Ethiopian origin, has a widespread distribution across arid, semi-arid and Mediterranean parts of the Arabian sub-region. We compared the daily energy expenditure (DEE), water turnover NTTO) and sustained metabolic scope (SusMS = DEE/resting metabolic rate) of two adjacent populations during the winter. Mice were captured from North- and South- facing slopes (NFS and SFS) of the same valley, comprising mesic and xeric habitats, respectively. Both DEE and SusMS winter values were greater in NFS than SFS mice and were significantly greater than values previously measured in the summer for these two populations in the same environments. However, WTO values were consistent with previously established values and were not significantly different from allometric predictions for desert eutherians. We suggest that physiological plasticity in energy expenditure, which exists both temporally and spatially, combined with stable WTO, perhaps reflecting a xeric ancestry, has enabled A. cahirinus to invade a wide range of habitats. (C) 2003 Elsevier Inc. All rights reserved.

Uncertainty and Sensitivity Analysis in Aircraft Operating Costs in Structural Design Optimization

The paper focuses on the development of an aircraft design optimization methodology that models uncertainty and sensitivity analysis in the tradeoff between manufacturing cost, structural requirements, andaircraft direct operating cost.Specifically,ratherthanonlylooking atmanufacturingcost, direct operatingcost is also consideredintermsof the impact of weight on fuel burn, in addition to the acquisition cost to be borne by the operator. Ultimately, there is a tradeoff between driving design according to minimal weight and driving it according to reduced manufacturing cost. Theanalysis of cost is facilitated withagenetic-causal cost-modeling methodology,andthe structural analysis is driven by numerical expressions of appropriate failure modes that use ESDU International reference data. However, a key contribution of the paper is to investigate the modeling of uncertainty and to perform a sensitivity analysis to investigate the robustness of the optimization methodology. Stochastic distributions are used to characterize manufacturing cost distributions, andMonteCarlo analysis is performed in modeling the impact of uncertainty on the cost modeling. The results are then used in a sensitivity analysis that incorporates the optimization methodology. In addition to investigating manufacturing cost variance, the sensitivity of the optimization to fuel burn cost and structural loading are also investigated. It is found that the consideration of manufacturing cost does make an impact and results in a different optimal design configuration from that delivered by the minimal-weight method. However, it was shown that at lower applied loads there is a threshold fuel burn cost at which the optimization process needs to reduce weight, and this threshold decreases with increasing load. The new optimal solution results in lower direct operating cost with a predicted savings of 640=m2 of fuselage skin over the life, relating to a rough order-of-magnitude direct operating cost savings of $500,000 for the fuselage alone of a small regional jet. Moreover, it was found through the uncertainty analysis that the principle was not sensitive to cost variance, although the margins do change.

Costs, outputs and outcomes in school PFI contracts and the significance of project size

This article examines operational Private Finance Initiative (PFI) school projects and reports the experiences of UK headteachers. It considers the impact of project size on value for money (VFM). Headteachers involved in small projects are more satisfied with costs than those involved in large projects, but headteachers involved in larger projects are more satisfied with affordability. Generally, heads are more satisfied with the buildings than with the services. The authors question the government’s recent policy changes to increase the size of PFI projects.

A hierarchical multiclass support vector machine incorporated with holistic triple learning units

This paper proposes a new hierarchical learning structure, namely the holistic triple learning (HTL), for extending the binary support vector machine (SVM) to multi-classification problems. For an N-class problem, a HTL constructs a decision tree up to a depth of A leaf node of the decision tree is allowed to be placed with a holistic triple learning unit whose generalisation abilities are assessed and approved. Meanwhile, the remaining nodes in the decision tree each accommodate a standard binary SVM classifier. The holistic triple classifier is a regression model trained on three classes, whose training algorithm is originated from a recently proposed implementation technique, namely the least-squares support vector machine (LS-SVM). A major novelty with the holistic triple classifier is the reduced number of support vectors in the solution. For the resultant HTL-SVM, an upper bound of the generalisation error can be obtained. The time complexity of training the HTL-SVM is analysed, and is shown to be comparable to that of training the one-versus-one (1-vs.-1) SVM, particularly on small-scale datasets. Empirical studies show that the proposed HTL-SVM achieves competitive classification accuracy with a reduced number of support vectors compared to the popular 1-vs-1 alternative.

Assessing the Transaction Costs of Firms in the EU ETS: Lessons form Ireland

Until now, there has been little empirical evidence that EU Emissions Trading Scheme (ETS) transaction costs are incurred at firm level. The transaction costs (internal costs, capital costs, consultancy and trading costs) incurred by Irish firms under the EU ETS during its pilot phase (2005-2007) were measured and analysed. Evidence for the sources of transaction costs, their magnitude and the distribution of costs shows that these were mainly administrative in nature. Considerable variation in costs was found due to economies of scale, as the costs per tonne of CO2 were lower for participants with larger allocations. For the largest firms - accounting for over half the emissions - average transaction costs were €0.05 per tonne. However, for small firms, average transaction costs were €2.02 - over 18% of the current allowance price. This supports the concerns that transaction costs are excessive for smaller participants. The immediate policy implication is that additional attention will be needed to address different sizes of firms, number of installations per firm, and the size of the initial allocations.