Showing papers by "University of New Brunswick published in 2004"

The Optimality of Naive Bayes.

Abstract: Naive Bayes is one of the most efficient and effective inductive learning algorithms for machine learning and data mining. Its competitive performance in classification is surprising, because the conditional independence assumption on which it is based, is rarely true in realworld applications. An open question is: what is the true reason for the surprisingly good performance of naive Bayes in classification? In this paper, we propose a novel explanation on the superb classification performance of naive Bayes. We show that, essentially, the dependence distribution; i.e., how the local dependence of a node distributes in each class, evenly or unevenly, and how the local dependencies of all nodes work together, consistently (supporting a certain classification) or inconsistently (canceling each other out), plays a crucial role. Therefore, no matter how strong the dependences among attributes are, naive Bayes can still be optimal if the dependences distribute evenly in classes, or if the dependences cancel each other out. We propose and prove a sufficient and necessary conditions for the optimality of naive Bayes. Further, we investigate the optimality of naive Bayes under the Gaussian distribution. We present and prove a sufficient condition for the optimality of naive Bayes, in which the dependence between attributes do exist. This provides evidence that dependence among attributes may cancel out each other. In addition, we explore when naive Bayes works well. Naive Bayes and Augmented Naive Bayes Classification is a fundamental issue in machine learning and data mining. In classification, the goal of a learning algorithm is to construct a classifier given a set of training examples with class labels. Typically, an example E is represented by a tuple of attribute values (x1, x2, , · · · , xn), where xi is the value of attribute Xi. Let C represent the classification variable, and let c be the value of C. In this paper, we assume that there are only two classes: + (the positive class) or − (the negative class). A classifier is a function that assigns a class label to an example. From the probability perspective, according to Bayes Copyright c © 2004, American Association for Artificial Intelligence (www.aaai.org). All rights reserved. Rule, the probability of an example E = (x1, x2, · · · , xn) being class c is p(c|E) = p(E|c)p(c) p(E) . E is classified as the class C = + if and only if fb(E) = p(C = +|E) p(C = −|E) ≥ 1, (1) where fb(E) is called a Bayesian classifier. Assume that all attributes are independent given the value of the class variable; that is, p(E|c) = p(x1, x2, · · · , xn|c) = n ∏

Topologies of single-phase inverters for small distributed power generators: an overview

Abstract: This paper presents an overview of single-phase inverters developed for small distributed power generators. The functions of inverters in distributed power generation (DG) systems include dc-ac conversion, output power quality assurance, various protection mechanisms, and system controls. Unique requirements for small distributed power generation systems include low cost, high efficiency and tolerance for an extremely wide range of input voltage variations. These requirements have driven the inverter development toward simpler topologies and structures, lower component counts, and tighter modular design. Both single-stage and multiple-stage inverters have been developed for power conversion in DG systems. Single-stage inverters offer simple structure and low cost, but suffer from a limited range of input voltage variations and are often characterized by compromised system performance. On the other hand, multiple-stage inverters accept a wide range of input voltage variations, but suffer from high cost, complicated structure and low efficiency. Various circuit topologies are presented, compared, and evaluated against the requirements of power decoupling and dual-grounding, the capabilities for grid-connected or/and stand-alone operations, and specific DG applications in this paper, along with the identification of recent development trends of single-phase inverters for distributed power generators.

An intelligent maximum power extraction algorithm for inverter-based variable speed wind turbine systems

Abstract: This paper focuses on the development of maximum wind power extraction algorithms for inverter-based variable speed wind power generation systems. A review of existing maximum wind power extraction algorithms is presented in this paper, based on which an intelligent maximum power extraction algorithm is developed by the authors to improve the system performance and to facilitate the control implementation. As an integral part of the max-power extraction algorithm, advanced hill-climb searching method has been developed to take into account the wind turbine inertia. The intelligent memory method with an on-line training process is described in this paper. The developed maximum wind power extraction algorithm has the capability of providing initial power demand based on error driven control, searching for the maximum wind turbine power at variable wind speeds, constructing an intelligent memory, and applying the intelligent memory data to control the inverter for maximum wind power extraction, without the need for either knowledge of wind turbine characteristics or the measurements of mechanical quantities such as wind speed and turbine rotor speed. System simulation results and test results have confirmed the functionality and performance of this method.

Agostic Interactions in d0 Metal Alkyl Complexes

Abstract: The phenomenon of agostic interactions is reviewed and the nature of the interaction is revisited. A historical perspective is followed by an overview of experimental techniques used to diagnose agostic behavior, and previous interpretations of agostic bonding are presented. A series of simple metal alkyl complexes is considered and a new model for the phenomenon in d(0) systems is developed which sets them apart from agostic late-transition-metal complexes. Factors such as the valence electron count and coordination number of the metal center are revealed to be unimportant in facilitating the interaction in most d(0) systems. The charge density distribution in several transition-metal alkyl complexes is explored by experimental and theoretical techniques, including the powerful "Atoms in Molecules" approach. Local charge concentrations are shown to play an important role in the agostic interaction. Finally, we demonstrate for the first time a way to manipulate and control the magnitude and disposition of such local charge concentrations, and hence the strength of agostic interactions in d(0) metal alkyl complexes.

Modelling strategies for controlling SARS outbreaks

Abstract: Severe acute respiratory syndrome (SARS), a new, highly contagious, viral disease, emerged in China late in 2002 and quickly spread to 32 countries and regions causing in excess of 774 deaths and 8098 infections worldwide. In the absence of a rapid diagnostic test, therapy or vaccine, isolation of individuals diagnosed with SARS and quarantine of individuals feared exposed to SARS virus were used to control the spread of infection. We examine mathematically the impact of isolation and quarantine on the control of SARS during the outbreaks in Toronto, Hong Kong, Singapore and Beijing using a deterministic model that closely mimics the data for cumulative infected cases and SARS-related deaths in the first three regions but not in Beijing until mid-April, when China started to report data more accurately. The results reveal that achieving a reduction in the contact rate between susceptible and diseased individuals by isolating the latter is a critically important strategy that can control SARS outbreaks with or without quarantine. An optimal isolation programme entails timely implementation under stringent hygienic precautions defined by a critical threshold value. Values below this threshold lead to control, but those above are associated with the incidence of new community outbreaks or nosocomial infections, a known cause for the spread of SARS in each region. Allocation of resources to implement optimal isolation is more effective than to implement sub-optimal isolation and quarantine together. A community-wide eradication of SARS is feasible if optimal isolation is combined with a highly effective screening programme at the points of entry.

The relationship between maximal jump-squat power and sprint acceleration in athletes

Abstract: This study investigated the relationship between sprint start performance (5-m time) and strength and power variables. Thirty male athletes [height: 183.8 (6.8) cm, and mass: 90.6 (9.3) kg; mean (SD)] each completed six 10-m sprints from a standing start. Sprint times were recorded using a tethered running system and the force-time characteristics of the first ground contact were recorded using a recessed force plate. Three to six days later subjects completed three concentric jump squats, using a traditional and split technique, at a range of external loads from 30–70% of one repetition maximum (1RM). Mean (SD) braking impulse during acceleration was negligible [0.009 (0.007) N/s/kg) and showed no relationship with 5 m time; however, propulsive impulse was substantial [0.928 (0.102) N/s/kg] and significantly related to 5-m time (r=−0.64, P<0.001). Average and peak power were similar during the split squat [7.32 (1.34) and 17.10 (3.15) W/kg] and the traditional squat [7.07 (1.25) and 17.58 (2.85) W/kg], and both were significantly related to 5-m time (r=−0.64 to −0.68, P<0.001). Average power was maximal at all loads between 30% and 60% of 1RM for both squats. Split squat peak power was also maximal between 30% and 60% of 1RM; however, traditional squat peak power was maximal between 50% and 70% of 1RM. Concentric force development is critical to sprint start performance and accordingly maximal concentric jump power is related to sprint acceleration.

Measurement and Validity Characteristics of the Short Version of the Social and Emotional Loneliness Scale for Adults

Abstract: This article presents a psychometric study of the short form of the Social and Emotional Loneliness Scale for Adults (SELSA-S). Data were collected via self-report measures and mail surveys from se...

Passive hyperthermia reduces voluntary activation and isometric force production

Abstract: It has been suggested that a critically high body core temperature may impair central neuromuscular activation and cause fatigue. We investigated the effects of passive hyperthermia on maximal isometric force production (MVC) and voluntary activation (VA) to determine the relative roles of skin (T sk) and body core temperature (T c) on these factors. Twenty-two males [VO2max=64.2 (8.9) ml kg−1 min−1, body fat=8.2 (3.9)%] were seated in a knee-extension myograph, then passively heated from 37.4 to 39.4°C rectal temperature (T re) and then cooled back to 37.4oC using a liquid conditioning garment. Voluntary strength and VA (interpolated twitch) were examined during an isometric 10-s MVC at 0.5°C intervals during both heating and cooling. Passive heating to a T c of 39.4oC reduced VA by 11 (11)% and MVC by 13 (18)% (P<0.05), but rapid skin cooling, with a concomitant reduction in cardiovascular strain [percentage heart rate reserve decreased from 64 (11)% to 29 (11)%] and psychophysical strain did not restore either of these measures to baseline. Only when cooling lowered T c back to normal did VA and MVC return to baseline (P<0.05). We conclude that an elevated T c reduces VA during isometric MVC, and neither T sk nor cardiovascular or psychophysical strain modulates this response. Results are given as mean (SD) unless otherwise stated.

A Picture and 1000 Words: Using Resident-Employed Photography to Understand Attachment to High Amenity Places

Abstract: Research on attachment to high amenity places has usually focused on visitors, despite the fact that many of these settings also may hold permanent residents. Visitor employed photography (VEP) has...

Harmonic radar transceiver design: miniature tags for insect tracking

Abstract: The design and operation along with verifying measurements of a harmonic radar transceiver, or tag, developed for insect tracking are presented. A short length of wire formed the antenna while a beam lead Schottky diode across a resonant loop formed the frequency doubler circuit yielding a total tag mass of less than 3 mg. Simulators using the method-of-moments for the antenna, finite-integral time-domain for the loop, and harmonic balance for the nonlinear diode element were used to predict and optimize the transceiver performance. This performance is compared to the ideal case and to measurements performed using a pulsed magnetron source within an anechoic chamber. A method for analysis of the tag is presented and used to optimize the design by creating the largest possible return signal at the second harmonic frequency for a particular incident power density. These methods were verified through measurement of tags both in isolation and mounted on insects. For excitation at 9.41 GHz the optimum tag in isolation had an antenna length of 12 mm with a loop diameter of 1 mm which yielded a harmonic cross-section of 40 mm/sup 2/. For tags mounted on Colorado potato beetles, optimum performance was achieved with an 8 mm dipole fed 2 mm from the beetle attached end. A theory is developed that describes harmonic radar in a fashion similar to the conventional radar range equation but with harmonic cross-section replacing the conventional radar cross-section. This method provides a straightforward description of harmonic radar system performance as well as provides a means to describe harmonic radar tag performance.

Perspectives on Authenticity in Teaching

Abstract: The authors work with 22 educators from a variety of disciplines during a 3-year time span to understand what authentic teaching means and to explore how authenticity is manifested in practice. Usi...

Development of a novel wind turbine simulator for wind energy conversion systems using an inverter-controlled induction motor

Abstract: A wind turbine simulator for wind energy conversion systems has been developed with a view to design, evaluate, and test of actual wind turbine drive trains including generators, transmissions, power-electronic converters and controllers. The simulator consists of a 10-hp induction motor (IM) which drives a generator and is driven by a 10-kW variable speed drive inverter and real-time control software. In this simulator, a microcontroller, a PC interfaced to LAB Windows I/O board, and an IGBT inverter-controlled induction motor are used instead of a real wind turbine to supply shaft torque. A control program based on C language is developed that obtains wind profiles and, by using turbine characteristics and rotation speed of IM, calculates the theoretical shaft torque of a real wind turbine. Comparing with this torque value, the shaft torque of the IM is regulated accordingly by controlling stator current demand and frequency demand of the inverter. In this way, the inverter driven induction motor acts like a real wind turbine to the energy conversion system. The drive is controlled using the measured shaft torque directly, instead of estimating it as conventional drives do. The experimental results of the proposed simulator show that this scheme is viable and accurate. This paper reports the operating principles, theoretical analyses, and test results of this wind turbine simulator.

Psychopathic traits in adolescent offenders: an evaluation of criminal history, clinical, and psychosocial correlates.

Abstract: Although a large body of research has established the relevance of psychopathy to adult offenders, its relevance to adolescent offenders is far less clear. The current study evaluated the clinical, psychosocial and criminal correlates of psychopathic traits in a sample of 226 male and female incarcerated adolescent offenders. According to an 18-item version of the Psychopathy Checklist-Youth Version (PCL-YV; Forth, Kosson, & Hare, 2003), only 9.4% exhibited a high level of psychopathic traits (PCL-YV>/=25). Consistent with past research, higher PCL-YV scores were positively associated with self-reported delinquency and aggressive behavior and were unrelated to emotional difficulties. Although higher PCL-YV scores were associated with the experience of physical abuse, the only psychosocial factor to predict PCL-YV scores was a history of non-parental living arrangements (e.g. foster care). In terms of criminality, a violent/versatile criminal history was positively associated with psychopathic traits. However, PCL-YV scores were unrelated to participants' official criminal records for total, non-violent, violent, and technical violation convictions. In conclusion, the data partially support the construct validity of psychopathy with adolescent offenders, but some inconsistencies with prior adult and adolescent psychopathy research were evident. These issues are discussed.

Reversible cascades with minimal garbage

Abstract: The problem of minimizing the number of garbage outputs is an important issue in reversible logic design. We start with the analysis of the number of garbage outputs that must be added to a multiple output function to make it reversible. We give a precise formula for the theoretical minimum of the required number of garbage outputs. For some benchmark functions, we calculate the garbage required by some proposed reversible design methods and compare it to the theoretical minimum. Based on the information about minimal garbage, we suggest a new reversible design method that uses the minimum number of garbage outputs. We show that any Boolean function can be realized as a reversible network in terms of this new approach by giving the theoretical method of finding such a network. Using a heuristics synthesis approach, we create a program and run it to compare results of our synthesis to the previously reported synthesis results for the benchmark functions with up to ten variables. Finally, we show that the synthesis for the proposed model can be accomplished with lower cost than the synthesis of EXOR programmable logic arrays.

Properties of furfurylated wood

Abstract: The first processes of wood modification with furfuryl alcohol (FA) (furfurylation) were developed several decades ago. FA is a renewable chemical, produced from hydrolysed biomass waste. Over the past decade modernized processes for furfurylation of wood have been developed. This study presents decay properties of furfurylated wood. Laboratory methods and field tests were performed on fungi, termite and marine borer attack. Tests on physical and mechanical properties are also presented. The properties of furfurylated wood depend on the retention of grafted/polymerized poly-FA in the wood. At high modification levels (high retention of poly-FA) the enhancement of a wide variety of properties is achieved: an exceptional increase in hardness, exceptional resistance to microbial decay and insect attack, increase in modulus of rupture (MOR) and modulus of elasticity, and high dimensional stability. At lower modification levels property enhancements also occur. Notable are resistance to microbial decay and ins...

Estimating the ‘critical’ distance at which adjacent land-use degrades wetland water and sediment quality

Abstract: Conversion of forested lands to agriculture or urban/residential areas has been associated with declines in stream and lake water quality. Less attention has been paid to the effects of adjacent land-uses on wetland sediment and water quality and, perhaps more importantly, the spatial scales at which these effects occur. Here we address these issues by examining variation in water and sediment nutrient levels in 73 southeastern Ontario, Canada, wetlands. We modeled the relationship between water and sediment nutrient concentrations and various measures of adjacent land-use such as forest cover and road density, measured over increasing distances from the wetland edge. We found that water nitrogen and phosphorous levels were negatively correlated with forest cover at 2250 meters from the wetland edge, while sediment phosphorous levels were negatively correlated with wetland size and forest cover at 4000 meters and positively correlated with the proportion of land within 4000 meters that is itself wetland. These results suggest that the effects of adjacent land-use on wetland sediment and water quality can extend over comparatively large distances. As such, effective wetland conservation will not be achieved merely through the creation of narrow buffer zones between wetlands and more intensive land-uses. Rather, sustaining high wetland water quality will require maintaining a heterogeneous regional landscape containing relatively large areas of natural forest and wetlands.

Development of a psychometrically sound internet addiction scale: a preliminary step.

Abstract: The Internet Addiction Scale (IAS) is a self-report instrument based on the 7 Diagnostic and Statistical Manual of Mental Disorders (4th ed.; American Psychiatric Association, 1994) substance dependence criteria and 2 additional criteria recommended by Griffiths (1998). The IAS was administered to 233 undergraduates along with 4 measures pertaining to loneliness and boredom proneness. An item reliability analysis reduced the initial scale from 36 to 31 items (with a Cronbach's alpha of .95). A principal-components analysis indicated that the IAS consisted mainly of one factor. Multiple regression analyses revealed that Family and Social Loneliness and Boredom Proneness were significantly correlated with the IAS; Family and Social Loneliness uniquely predicted IAS scores. No evidence for widespread Internet addiction was found.

The igneous rocks of Greece: The anatomy of an orogen. Georgia Pe-Piper and David J.W. Piper.

Abstract: The Hellenide orogen in Greece is part of the AlpineHimalaya mountian belt, created during the destruction of Tethys by the convergence of Gondwana and Eurasia. Within Greece, there is the record of a complete Wilson tectonic cycle of continental rifting, sea-fl oor spreading, plate subduction, and continental collision during Mesozoic and Cenozoic time. This book presents a new synthesis of the geological history of Greece as revealed by the varied igneous rocks. It is based on more than 30 years of fi eld and laboratory studies by the authors together with a synthesis of the widely scattered published literature that has been written in many different languagues. Basement rocks record Hercynian subduction and plutonism on the northern margin of Gondwana, which in the Permian and Triassic rifted into several microcontinents, thereby creating the eastern Mediterranean Neotethys ocean. Partial closure of strands of the Neotethys ocean resulted in widespread emplacement of Jurassic and Cretaceous ophiolites. Early Tertiary collision produced a Hellenide mountian chain similar to the Alps and Himalayas. Rapid Neogene extension of the Hellenides behind the modern South Aegean arc has formed the Aegean Sea, triggered widespread backarc igneous activity, and unroofed mid-crustal rocks. The geological setting, geochemistry, and tectonic signifi cance of each group of rocks is presented in detail, with numerous original maps and fi gures.

Assessing red algal supraordinal diversity and taxonomy in the context of contemporary systematic data

Abstract: The wondrously diverse eukaryotes that constitute the red algae have been the focus of numerous recent molecular surveys and remain a rich source of undescribed and little known species for the traditional taxonomist. Molecular studies place the red algae in the kingdom Plantae; however, supraordinal classification has been largely confined to debate on subclass vs. class level status for the two recognized subgroups, one of which is widely acknowledged as paraphyletic. This narrow focus has generally masked the extent to which red algal classification needs modification. We provide a comprehensive review of the literature pertaining to the antiquity, diversity, and systematics of the red algae and propose a contemporary classification based on recent and traditional evidence.

Performance of pfa concrete in a marine environment––10-year results

Abstract: Steel reinforced 100×100×300 mm concrete prisms, with nominal strength grades C25, C35 and C45 and different pfa levels (0–50%), were exposed to various curing treatments during the first 28 days prior to exposure in the tidal zone of the BRE marine exposure site. Chloride concentration profiles and rebar weight losses had previously been measured on specimens exposed for 1, 2 and 4 years and this paper reports the results of similar measurements after 10 years exposure. Chloride profiles were also measured for specimens after 1 and 28 days immersion in seawater under laboratory conditions. Pfa concretes showed substantially increased resistance to the penetration of chlorides compared with control Portland cement (PC) concrete specimens. The improved resistance of the pfa concrete to the penetration of chlorides resulted in reduced corrosion of steel bars imbedded in the concrete. Threshold chloride levels for corrosion, estimated from relationships between steel weight loss and chloride content at the location of the steel, were found to decrease with increasing pfa content. Chloride concentration profiles after 28 days of immersion in seawater showed that considerable chloride penetration occurred during this period due to sorption (capillary suction) of the seawater into the unsaturated specimens. This results in a significant error in diffusion coefficients calculated from the concentration profile using the standard solution to Fick’s second law. The error may be substantial for pfa concretes where chlorides penetrating due to sorption immediately after exposure may outweigh subsequent diffusion during continued seawater exposure. The performance of the concretes, particularly the PC concretes, in this programme is considered in the light of current and new British Standard recommendations for concrete exposed to marine tidal conditions. The adequacy of these recommendations is, however, difficult to assess because the highest concrete quality tested fell a little short of the minimum quality required in the recommendations and cover to reinforcement was also less than required. Nevertheless, the superior performance of concretes containing 30% or more pfa was clearly demonstrated.

Effects of clutter on echolocation call structure of myotis septentrionalis and m. lucifugus

Abstract: The structure of echolocation calls, and the distance over which bats perceive their environment, varies with the amount of structural clutter through which they are flying. Clutter and species had significant effects on the frequency-time characteristics of search-phase echolocation calls of northern long-eared (Myotis septentrionalis) and little brown bats (M. lucifugus). We tested an a priori derived model that predicted the pattern of differences in echolocation call variable values among clutter categories would provide insight into the relative maximum distances that bat species could perceive using echolocation. Specifically, the model predicted that species adapted to flying and foraging in cluttered habitats would have a shorter maximum perceptual distance than species adapted to flying and foraging in uncluttered habitats. The results supported this model and suggest the clutter-adapted M. septentrionalis had a shorter maximum perceptual distance than M. lucifugus, a species known to forage in a variety of habitats but mainly in uncluttered areas (i.e., over water). Using calls as the sampling unit, a neural network correctly classified >94% of the echolocation calls to species in high clutter. In medium and low clutter, >82% of the calls were correctly classified to species; however >90% correct classification was achieved by leaving <30% of calls unclassified. Researchers should develop clutter-specific call libraries to improve species classification accuracy for echolocation calls.

Resource dependence and community well-being in rural Canada.

Abstract: The well-being of residents of resource dependent communities is a question of traditional interest to rural sociologists. The label “resource dependent” obscures how this relationship may vary between particular resource industries, regions, or indicators of well-being. Few analyses have compared the relationship between well-being and resource dependence across different industries, nor tested competing theories about the relationship between resource dependence and well-being. Our paper presents an overview of the relationship between resource dependence—agriculture, fisheries, mining, energy, forestry—and human well-being in Canada. Analysis of 1996 Statistics Canada data revealed a great deal of variation in the effect of “resource” dependence on indicators of well-being (e.g., human capital, unemployment, income): some industries exhibit fairly positive outcomes (e.g., agriculture), others more negative outcomes (e.g., fishing). Consistent with analyses conducted in the United States, these relationships vary by region, suggesting the need for models that incorporate the particulars of place and industry.

Density‐dependent habitat selection and the ideal free distribution in marine fish spatial dynamics: considerations and cautions

Abstract: Current methods and theory used in the study of the spatial dynamics of marine fish are problematic. Positive relationships between population abundance and occupied area are typically interpreted as evidence of density-dependent habitat selection. However, both abundance and area may co-vary with an un-parameterized variable, such as a density-independent effect. In addition, if density-dependent habitat selection is present, population growth rates in optimal habitats would be expected to be lower than in marginal habitats. This same pattern can also evolve from a largescale, spatially autocorrelated change in a density-independent factor. The theory underlying density-dependent habitat selection, the ideal free distribution, can be tautological when no a priori information of how habitat suitability changes with density is known. In this case, an ideal free distribution can be defined for any pattern of habitat-specific population growth rates. However, these problems are not insurmountable and solutions may be found by considering spatial variation in proxies of fitness and explicitly allowing for the relative importance of habitat selection (density dependent) and environmental (density independent) effects to vary with spatial scale.