On the evaluation of unsupervised outlier detection: measures,datasets, and an empirical study

The evaluation of unsupervised outlier detection algorithms is a constant challenge in data mining research. Little is known regarding the strengths and weaknesses of different standard outlier detection models, and the impact of parameter choices for these algorithms. The scarcity of appropriate benchmark datasets with ground truth annotation is a significant impediment to the evaluation of outlier methods. Even when labeled datasets are available, their suitability for the outlier detection task is typically unknown. Furthermore, the biases of commonly-used evaluation measures are not fully understood. It is thus difficult to ascertain the extent to which newly-proposed outlier detection methods improve over established methods. In this paper, we perform an extensive experimental study on the performance of a representative set of standard k nearest neighborhood-based methods for unsupervised outlier detection, across a wide variety of datasets prepared for this purpose. Based on the overall performance of the outlier detection methods, we provide a characterization of the datasets themselves, and discuss their suitability as outlier detection benchmark sets. We also examine the most commonly-used measures for comparing the performance of different methods, and suggest adaptations that are more suitable for the evaluation of outlier detection results.     

Serotonin 5-HT2 receptors in schizophrenia: a PET study using [18F]setoperone in neuroleptic-naive patients and normal subjects

BACKGROUND: In the FRISC trial, dalteparin 120 IU/kg body weight twice daily for unstable coronary artery disease was safe and reduced the risk of new coronary events. This risk reduction was maintained during the following extended treatment with a fixed dose of 7500 IU dalteparin once daily. METHODS AND RESULTS: Minor bleeding was more frequent in women compared with men: relative risk (CI) 2.88 (1.78 to 4.67) during the weight-adjusted and 2.36 (1.37 to 2.63) during the fixed dose treatment. The anti-Xa activity determined in samples (n = 175) obtained during the acute phase treatment was higher in women compared with men (P <.001) and in nonsmokers compared with smokers (<.001) in multiple regression analysis. Also, during the fixed-dose treatment (n = 131) an independent relation between anti-Xa activity and sex (P <.001), but not smoking habits, persisted. CONCLUSION: To improve future low-molecular-weight heparin dose regimens for the treatment of acute coronary syndromes, it might be important to consider the influence of sex and smoking habits.     

Robust Distance-Based Clustering with Applications to Spatial Data Mining

In this paper we present a method for clustering geo-referenced data suitable for applications in spatial data mining, based on the medoid method. The medoid method is related to k -MEANS, with the restriction that cluster representatives be chosen from among the data elements. Although the medoid method in general produces clusters of high quality, especially in the presence of noise, it is often criticized for the Ω(n ² ) time that it requires. Our method incorporates both proximity and density information to achieve high-quality clusters in subquadratic time; it does not require that the user specify the number of clusters in advance. The time bound is achieved by means of a fast approximation to the medoid objective function, using Delaunay triangulations to store proximity information. Received December 21, 1998; revised August 25, 1999, and October 25, 1999.     

Characterization of the temperature sensitivity of gain and recombination mechanisms in 1.3-/spl mu/m AlGaInAs MQW lasers

The potential of 1.3-/spl mu/m AlGaInAs multiple quantum-well (MQW) laser diodes for uncooled operation in high-speed optical communication systems is experimentally evaluated by characterizing the temperature dependence of key parameters such as the threshold current, transparency current density, optical gain and carrier lifetime. Detailed measurements performed in the 20/spl deg/C-100/spl deg/C temperature range indicate a localized T/sub 0/ value of 68 K at 98/spl deg/C for a device with a 2.8 /spl mu/m ridge width and 700-/spl mu/m cavity length. The transparency current density is measured for temperatures from 20/spl deg/C to 60/spl deg/C and found to increase at a rate of 7.7 A/spl middot/cm/sup -2//spl middot/ /spl deg/C/sup -1/. Optical gain characterizations show that the peak modal gain at threshold is independent of temperature, whereas the differential gain decreases linearly with temperature at a rate of 3/spl times/10/sup -4/ A/sup -1//spl middot//spl deg/C/sup -1/. The differential carrier lifetime is determined from electrical impedance measurements and found to decrease with temperature. From the measured carrier lifetime we derive the monomolecular ( A), radiative (B), and nonradiative Auger (C) recombination coefficients and determine their temperature dependence in the 20/spl deg/C-80/spl deg/C range. Our study shows that A is temperature independent, B decreases with temperature, and C exhibits a less pronounced increase with temperature. The experimental observations are discussed and compared with theoretical predictions and measurements performed on other material systems.     

Modeling with blocks

This paper presents a simple and general modeling primitive, called a block, based on a generalized cuboid shape. Blocks are laid out and connected together to constitute the base shape of complex objects, from which is extracted a control mesh that can contain both smooth and sharp edges. The volumetric nature of the blocks allows for easy topology specification, as well as CSG operations between blocks. The surface parameterization inherited from the block faces provides support for texturing and displacement functions to apply surface details. A?variety of examples illustrate the generality of our blocks in both interactive and procedural modeling contexts.     

Computer aided evaluation of pipeline current near toroidal HVDCground electrodes

The effect of the return current of an HVDC toroidal electrode on underground metallic installations such as pipelines and cables is evaluated. In order to quantify the effect of HVDC electrodes on the buried conductor, an equivalent model is first presented that takes into account the electrical properties of the pipeline and its surrounding. In particular, infinitely long pipelines with different coatings are considered. The pipelines are assumed to be continuously buried in homogeneous soil and are located in the vicinity of the HVDC electrode. It is found that the voltage gradient in the vicinity of the electrode results in return currents in grounded conductors, the effects of which can extend several kilometers. Protective coatings reduce stray current from local voltage disturbances but have limited effect against distant electrodes     

18F]fluoroalkyl analogues of the potent 5-HT1A antagonist WAY 100635: radiosyntheses and in vivo evaluation

Two analogues of the potent 5-HT1A antagonist WAY 100635 have been synthesized and radiolabelled with 18F, namely N-[2-[4-(2-2'-[18F] fluoroethoxyphenyl)-1-piperazinyl]ethyl]-N-(2-pyridinyl)cyclohe xan e carboxamide ([18F]FEC) and N-[2-[4-(2-3'-[18F] fluoropropoxyphenyl)-1-piperazinyl]ethyl]-N-(2-pyridinyl)cycloh exa ne carboxamide ([18F]FPC). Biodistribution studies in rats showed selective uptake of both radiotracers in regions known to be rich in 5-HT1A receptors following i.v. injection. The ratio of radioactivity in hippocampus to that in the cerebellum was 5.5 (for [18F]FEC) and 7.5 (for [18F]FPC) at 60 min postinjection. Regional brain heterogeneity of radioactivity could be abolished by pretreatment with WAY 100635 and FPC but was unaffected by pretreatment with a variety of drugs including ketanserin, sulpiride, and SCH 23390. These results are compared vis-a-vis with those obtained using [11C]WAY 100635 to evaluate [18F]FEC and [18F]FPC as potential radiotracers for imaging 5-HT1A receptors by positron emission tomography.     

Conceptual design and cost study for a dual-purpose nuclear-electric reverse osmosis seawater conversion plant

This paper presents the results of a study to develop a conceptual design, and cost estimate for a 25 million gallon per day seawater reverse osmosis desalting plant operating at both Caribbean and Persian Gulf sites. The plants operate in conjunction with a 1000 MWe nuclear power plant. Four seawater membrane manufacturers were supplied with feedwater analyses and a simplified cost estimating procedure so that they could recommend an optimum membrane system. From this information, plant designs and cost estimates were developed. For both sites, a two-stage system was selected for the conceptual cost estimate. The product water cost was determined, based- upon 1978 construction costs, for both the Caribbean, and Persian Gulf sites. In addition, areas of potential cost reduction were discussed.     

分布式环境下多关键词并行密文检索方案

对于可搜索加密需要均衡数据的安全性和检索效率。针对SSE-1密文检索方案中检索性能低、单关键词检索模式不足和传统单服务器架构中的单机资源局限性等问题，设计并实现了一种多关键词并行密文检索系统。该系统采用不同的索引加密方式提高密文检索性能；通过对密文倒排索引的切分实现倒排索引的分块检索，克服了单机资源的局限性并提高了检索效率；通过结合分布式特点扩展了传统单机检索架构并实现了多关键词的并行检索。实验结果表明，与SSE-1方案相比，在保证密文数据安全性的前提下所提方案能够提高检索、更新等操作的效率，实现多关键词的检索，同时动态扩展系统分布式架构以提高系统负载能力。