Thesis (PhD (Archaeology))--University of Stellenbosch, 2005. / This study investigates the evolution of modern mammalian faunas in the central interior of southern Africa by testing the hypothesis that the evolution of the black wildebeest, Connochaetes gnou, was directly associated with the emergence of Highveld-type open grasslands in the central interior. Southern Africa can be distinguished from other arid and semi-arid parts of the continent by the presence of an alliance of endemic grazing ungulates. The black wildebeest is characteristic of this alliance. Open habitats are essential for the reproductive behaviour of the black wildebeest, because territorial males require an unobstructed view of their territories in order to breed. The specialised territorial breeding behaviour of the black wildebeest is the reason why the black wildebeest is historically confined to the Highveld and Karoo areas and why it is reproductively isolated from sympatric blue wildebeest, Connochaetes taurinus. The finds from a number of fossil-rich localities, dating from the recent past to approximately a million years ago, have been identified. The remains referred to ancestral C. gnou have been subjected to detailed qualitative and quantitative osteological comparisons with cranial and post-cranial elements of modern and fossil reference specimens. This material includes extant southern African alcelaphines and fossil materials of C. gnou, the extinct giant wildebeest, Megalotragus priscus, and North African fossil alcelaphines. The results show that cranial changes in fossil C. gnou, particularly the more forward positioning of the horns, basal inflation of the horns and the resultant re-organisation of the posterior part of the skull, preceded other skeletal modifications. These cranial changes indicate a shift towards more specialised territorial breeding behaviour in the earliest ancestral black wildebeest, evident in the specimens of the c. million year old Free State site of Cornelia-Uitzoek. Since the territorial breeding behaviour of the black wildebeest can only function in open habitat and since cranial characters associated with its territorial breeding behaviour preceded other morphological changes, it is deduced that there was a close association between the speciation of C. gnou from a C. taurinus-like ancestor and the appearance of permanently open Highveld-type grasslands in the central interior of southern Africa. This deduction is supported by the lack of trophic distinction between the modern black and blue wildebeest, suggesting that the evolution of the black wildebeest was not accompanied by an ecological shift. It is concluded that the evolution of a distinct southern endemic wildebeest in the Pleistocene was associated with, and possibly driven by, a shift towards a more specialised kind of territorial breeding behaviour, which can only funtion in open habitat. There are significant post-speciation changes in body size and limb proportions of fossil C. gnou through time. The tempo of change has not been constant and populations in the central interior underwent marked reduction in body size in the last 5000 years. Vicariance in fossil C. gnou is evident in different rates of change that are recorded in the populations of generally smaller body size that became isolated in the Cape Ecozone. These daughter populations, the result of dispersals from the central interior, became extinct at the end of the Pleistocene.

Thesis (M.S.)--University of Tennessee, Knoxville, 2009. / Title from title page screen (viewed on Mar. 8, 2009). Thesis advisor: Craig Harper. Vita. Includes bibliographical references.

In this paper, a study is conducted to investigate the use of Curriculum Learning as an approach to address accuracy issues in a neural network caused by training on a Long-Tailed dataset. The thesis problem is presented by a Swedish e-commerce company. Currently, they are using a neural network that has been modified by them using a CORAL framework. This adaptation means that instead of having a classic binary regression model, it is an ordinal regression model. The data used for training the model has a Long-Tail distribution, which leads to inaccuracies when predicting a price distribution for items that are part of the tail-end of the data. The current method applied to remedy this problem is Re-balancing in the form of down-sampling and up-sampling. A linear training scheme is introduced, increasing in increments of $10\%$ while applying Curriculum Learning. As a method for sorting the data in an appropriate way, inspiration is drawn from Knowledge Distillation, specifically the Teacher-Student model approach. The teacher models are trained as specialists on three different subsets, and furthermore, those models are used as a basis for sorting the data before training the student model. During the training of the student model, the Curriculum Learning approach is used. The results show that for Imbalance Ratio, Kullback-Liebler divergence, Class Balance, and the Gini Coefficient, the data is clearly less Long-Tailed after dividing the data into subsets. With the correct settings before training, there is also an improvement in the training speed of the student model compared to the base model. The accuracy for both the student model and the base model is comparable. There is a slight advantage for the base model when predicting items in the head part of the data, while the student model shows improvements for items that are between the head and the tail. / I denna uppsats genomförs en studie för att undersöka användningen av Curriculum Learning som en metod för att hantera noggrannhetsproblem i ett neuralt nätverk som är en konsekvens av träning på data som har en Long-Tail fördelning. Problemstälnningen som behandlas i uppsatsen är tillhandagiven av ett svensk e-handelsföretag. För närvarande använder de ett neuralt nätverk som har modifierats med hjälp av ett CORAL-ramverk. Denna anpassning innebär att det istället för att ha en klassisk binär regressionsmodell har en ordinal regressionsmodell. Datan som används för att träna modellen har en Long-Tail fördelning, vilket leder till problem vid prediktering av prisfördelning för diverse föremål som tillhör datans svans. Den nuvarande metod som används för att åtgärda detta problem är en Re-balancing i form av down-sampling och up-sampling. Ett linjärt träningschema introduceras, som ökar i steg om $10\%$ medan Curriculum Learning tillämpas. Metoden för att sortera datan på ett lämpligt sätt inspires av Knowledge-Distillation, mer specifikt lärar-elevmodell delen. Lärarmodellerna tränas som specialister på tre olika delmängder, och därefter används dessa modeller som grund för att sortera datan innan tränandet av elevmodellen. Under träningen av elevmodellen tillämpas Curriculum Learning. Resultaten visar att för Imbalance Ratio, Kullback-Libler-divergens, Class Balance och Gini-koefficienten är datat tydligt mindre Long-Tailed efter att datat delats in i delmängder. Med rätt inställningar innan tränandet finns även en förbättring i träningshastighet för elevmodellen jämfört med basmodellen. Noggrannheten för både elevmodellen och basmodellen är jämförbar. Det finns en liten fördel för basmodellen vid prediktering av föremål i huvuddelen av datan, medan elevmodellen visar förbättringar för föremål som ligger mellan huvuddelen och svansen.

In den vergangenen Jahren ist die Untersuchung des Risikomanagements vom Baselkomitee angeregt, um die Kredit- und Bankwesen regelmäßig zu aufsichten. Für viele multivariate Risikomanagementmethoden gibt es jedoch Beschränkungen von: 1) verlässt sich die Kovarianzschätzung auf eine zeitunabhängige Form, 2) die Modelle beruhen auf eine unrealistischen Verteilungsannahme und 3) numerische Problem, die bei hochdimensionalen Daten auftreten. Es ist das primäre Ziel dieser Doktorarbeit, präzise und schnelle Methoden vorzuschlagen, die diesen Beschränkungen überwinden. Die Grundidee besteht darin, zuerst aus einer hochdimensionalen Zeitreihe die stochastisch unabhängigen Komponenten (IC) zu extrahieren und dann die Verteilungsparameter der resultierenden IC beruhend auf eindimensionale Heavy-Tailed Verteilungsannahme zu identifizieren. Genauer gesagt werden zwei lokale parametrische Methoden verwendet, um den Varianzprozess jeder IC zu schätzen, das lokale Moving Window Average (MVA) Methode und das lokale Exponential Smoothing (ES) Methode. Diese Schätzungen beruhen auf der realistischen Annahme, dass die IC Generalized Hyperbolic (GH) verteilt sind. Die Berechnung ist schneller und erreicht eine höhere Genauigkeit als viele bekannte Risikomanagementmethoden. / Over recent years, study on risk management has been prompted by the Basel committee for the requirement of regular banking supervisory. There are however limitations of many risk management methods: 1) covariance estimation relies on a time-invariant form, 2) models are based on unrealistic distributional assumption and 3) numerical problems appear when applied to high-dimensional portfolios. The primary aim of this dissertation is to propose adaptive methods that overcome these limitations and can accurately and fast measure risk exposures of multivariate portfolios. The basic idea is to first retrieve out of high-dimensional time series stochastically independent components (ICs) and then identify the distributional behavior of every resulting IC in univariate space. To be more specific, two local parametric approaches, local moving window average (MWA) method and local exponential smoothing (ES) method, are used to estimate the volatility process of every IC under the heavy-tailed distributional assumption, namely ICs are generalized hyperbolic (GH) distributed. By doing so, it speeds up the computation of risk measures and achieves much better accuracy than many popular risk management methods.

Value-at-Risk (VaR) model fails to predict financial risk accurately especially during financial crises. This is mainly due to the model’s inability to calibrate new market information and the fact that the risk measure is characterised by poor tail risk quantification. An alternative approach which comprises of the Expected Shortfall measure and the Lognormal Jump-Diffusion (LJD) model has been developed to address the aforementioned shortcomings of VaR. This model is called the Simulated-Expected-Shortfall (SES) model. The Maximum Likelihood Estimation (MLE) approach is used in determining the parameters of the LJD model since it’s more reliable and authenticable when compared to other nonconventional parameters estimation approaches mentioned in other literature studies. These parameters are then plugged into the LJD model, which is simulated multiple times in generating the new loss dataset used in the developed model. This SES model is statistically conservative when compared to peers which means it’s more reliable in predicting financial risk especially during a financial crisis. / Statistics / M.Sc. (Statistics)

Thesis (MSc)-- University of Stellenbosch, 1954 / Published in the Annals of the University of Stellenbosch, Volume 31, Section A, No.1 (1955) / ENGLISH ABSTRACT: The pelvic girdles of the two anuran genera Ascaphus and Leiopelma of the family Ascaphidae are very similar. Attached to them by cartilage (or connective tissue in some Ascaphus specimens) is a cartilaginous, in Leiopelmu somewhat ossified, epibubis. In Ascaphu8, as in Xenopus, the epipubis originates from two Anlages, and its muscles are also paired in the Ascaphidae, while in Xenopus only one is present, although this also appears to originate (rom two muscles. The cloaca extends behind the pelvic girdle in both sexes in Al:lcaphus, and it is supported by two rods consisting of strong connective tissue (Faserknochen?). These rods are attached to the ventral surface of the pelvic girdle, in the female closely, in the male by means of tendons. Between the rods and the epipubis a broad tendinous band extends, which is thick in the male; in the latter the mm. compressores cloacae have their origins posteriorly on the rods, while in the female the rods are completely imbedded in these muscles posteriorly. In the female particularly there are transverse muscle fibres which are not striped although probably derived from the mm. compressorcs cloacae. Behind the pelvic girdle there is cavernous tissue in both layers of the tunica muscularis, and ventrally it covers the mm. eompressores cloacae. The circular and longitudinal muscle layers are poste riorly separated by gliding planes. The cloaca is supplied by branches of the a. mcsenterica posterior and the au. pudendae anteriores, and is drained by the vv. pudendae and a small medial v. eaudalis. The cloaca is innervated dorsally by the plexus ischio·coccygeus and ventrally by a branch of the n.ischiadicus (the n. pudendus). The cloaca of the male serves as a copulatory organ. There is internal fertilization. / No Afrikaans abstract available. / The copy we received from the author were pages individually scanned in as JPEG images.

Master of Science / Department of Biology / Jack F. Cully, Jr. / Black-tailed prairie dogs (Cynomys ludovicianus) are a species of management and conservation concern. Prairie dogs have lost both habitat and occupied area due to plague, which is caused by the bacterium Yersinia pestis, pest control, and habitat conversion to agricultural land. Our goals were to estimate survival rates and dispersal rates, and to compare methods for estimating abundance of black-tailed prairie dogs for both management and conservation. We trapped black-tailed prairie dogs at four small National Parks from April 2009 through August 2011. Prairie dogs were trapped and marked for two trapping sessions per year in order to estimate seasonal rates of apparent survival. Apparent survival rates were estimated using the package RMark in R to construct models for program MARK. We found estimates to vary according to field site, sex, year, and season (summer or winter). Possible reasons for the differences in survivorship among sites could be presence of disease, quality of forage, predation, or frequency of dispersal. Visual counts were also conducted each trapping session beginning in April of 2010 to estimate abundance. Mark-recapture, mark-resight, and visual counts were compared to determine which method would be the most effective for estimating abundance of prairie dogs. We found mark-resight to produce the most precise estimates of abundance. While it costs more money to conduct a mark-resight estimate than visual counts because of repeated sessions, they produced significantly different results from one another 75% of the time, which was especially apparent on sites that had some form of visual barriers such as tall vegetation and uneven ground. However, if further information is needed in terms of sex ratios, age ratios, or the exact number of prairie dogs, then mark-recapture is the only method that can be used. Land managers need to address the level of accuracy needed, topography, and vegetation height before choosing which sampling method is best for the prairie dog towns in question. Finally, we looked at rates of intercolony and intracolony dispersal by placing 149 VHF collars and 6 GPS collars on prairie dogs at three colonies. Intracolony dispersal was also monitored through visual observation and trapping records over the three years of the study. We found 23 intracolony and eight intercolony dispersal events. Combined, these three studies offer insight not only into monitoring of prairie dog populations but also potential influence by plague both within and among colonies of prairie dogs.

Este trabalho tem como objetivo calcular intervalos de confiança para altos quantis oriundos de distribuições de caudas pesadas. Para isso, utilizamos os métodos da aproximação pela distribuição normal, razão de verossimilhanças, {\\it data tilting} e gama generalizada. Obtivemos, através de simulações, que os intervalos calculados a partir do método da gama generalizada apresentam probabilidades de cobertura bem próximas do nível de confiança, com amplitudes médias menores do que os outros três métodos, para dados gerados da distribuição Weibull. Todavia, para dados gerados da distribuição Fréchet, o método da razão de verossimilhanças fornece os melhores intervalos. Aplicamos os métodos utilizados neste trabalho a um conjunto de dados reais, referentes aos pagamentos de indenizações, em reais, de seguros de incêndio, de um determinado grupo de seguradoras no Brasil, no ano de 2003 / In this work, confidence intervals for high quantiles from heavy-tailed distributions were computed. More specifically, four methods, namely, normal approximation method, likelihood ratio method, data tilting method and generalised gamma method are used. A simulation study with data generated from Weibull distribution has shown that the generalised gamma method has better coverage probabilities with the smallest average length intervals. However, from data generated from Fréchet distribution, the likelihood ratio method gives the better intervals. Moreover, the methods used in this work are applied on a real data set from 1758 Brazilian fire claims

Este estudo apresenta dados sobre a ecologia alimentar do gavião-do-rabo-branco (Buteo albicaudatus albicaudatus) no município de Juiz de Fora, sudeste do estado de Minas Gerais. Os objetivos específicos foram: a) estudar quantitativamente a dieta do gavião-do-rabo-branco, incluindo número de presas e biomassa consumida; b) analisar possíveis padrões de sazonalidade na dieta; c) determinar a amplitude de nicho trófico do gavião-do-rabo-branco; d) verificar a existência de possíveis respostas numéricas e funcionais na dieta desse gavião e e) verificar estatisticamente as prévias proposições de oportunismo na escolha de presas por esse predador sugeridas por Stevenson & Meitzen (1946), Farquhar (1986) e Kopeny (1988). O material para estudo da dieta do B. albicaudatus consistiu de pelotas de regurgitação, coletado simultaneamente com os dados de abundância dos principais grupos de presas e do predador na área de estudo. As 259 amostras analisadas revelaram 31 espécies/morfoespécies e uma biomassa total estimada de 7196,5 g. Cinco ordens de insetos (Hymenoptera, Heteroptera, Odonata, Isoptera e Lepidoptera) e uma de vertebrado (Chiroptera) são descritas pela primeira vez como integrantes da dieta de B. albicaudatus. Numericamente, os invertebrados representaram 88% da dieta, sendo Acrididae responsável por 63% da dieta total. Dentre os vertebrados, os roedores foram o grupo mais consumido (73% da classe e 8% da dieta total). Analisando-se as presas em relação à biomassa, verificou-se uma brusca inversão na maioria das espécies/morfoespécies. Nesta abordagem, os vertebrados corresponderam a 77% da dieta. Os grupos mais representativos são: roedores (50%), ortópteros (21%), aves (11%), lagartos (7%) e serpentes (5%). A análise de sazonalidade na dieta mostrou uma evidente dependência de consumo quanto à estação climática, em mais da metade das categorias. Na estação seca verifica-se um maior consumo de Araneomorphae, outros artrópodes e mamíferos, enquanto na estação chuvosa, o maior consumo foi registrado para Acrididae e Coleoptera. Aves e répteis, apesar de não apresentarem uma dependência significativa em relação à sazonalidade, parecem ser utilizados como complemento ao consumo de mamíferos na estação chuvosa, no que tange a biomassa ingerida. Buteo albicaudatus pode ser classificado como insetívoro, em termos de número, ou como carnívoro, em termos de biomassa ingerida. A amplitude trófica da dieta mostrou-se de intermediária para baixa, variando de acordo com a abordagem adotada. Em termos gerais, quanto ao número e a biomassa, a dieta total mostrou que o gavião-do-rabo-branco é especialista. No entanto, a análise da biomassa referente a cada estação climática revelou que na estação chuvosa, a dieta é intermediária entre generalista e especialista (Bp itens=0,292; Bp grupos=0,485). Esses dados indicam que tanto o consumo de determinados itens alimentares quanto a amplitude de nicho trófico pode variar de acordo com a estação climática. Comparando-se a abundância do principal grupo de presa (pequenos mamíferos), em termos de biomassa, no ambiente e os registros de B. albicaudatus percebe-se que em nove dos doze meses há uma sincronia nas flutuações, caracterizando uma resposta numérica (rs=0,864; p<< 0,001). Resposta funcional foi registrada para a ordem Orthoptera (rs=0,762; p<0,01) e para a família Acrididae (rs=0,706; p<0,05). Com relação à seletividade de espécies de presas, observou-se que dos 12 gêneros de pequenos mamíferos capturados por armadilhas de interceptação e queda no ambiente, apenas cinco (Calomys, Akodon, Oligoryzomys, Oxymycterus e Gracilinanus) fazem parte da dieta de B. albicaudatus. Em relação a estes gêneros, o gavião-do-rabo-branco apresentou seletividade para Calomys tener e rejeição para Akodon spp apenas na estação seca. Os dados apresentados neste estudo sugerem que a abundância de pequenos mamíferos pode determinar o comportamento seletivo ou oportunismo de B. albicaudatus. No período de maior oferta de recurso há uma seletividade enquanto, no período de menor abundância de alimento, há um oportunismo por parte do predador. Este estudo revelou que o gavião-do-rabo-branco possui uma flexibilidade em sua dieta.

The Roatan Spiny-tailed Iguana (Ctenosaura oedirhina) is endemic to the 146-km2 island of Roatn, Honduras. Harvesting for consumption, fragmentation of habitat, and predation by domestic animals threaten this lizard. It is currently listed as Endangered by the International Union for Conservation of Nature (IUCN), as threatened by the Honduran government, and is on Appendix II of the Convention on International Trade in Endangered Species (CITES). This species has been geographically fragmented and genetically isolated into small subpopulations that are declining in density. With data gathered from use/availability surveys, resource selection functions were used to identify habitats and environmental variables associated with their presence. Results indicate that protection from harvesting is the most important factor in determining their distribution. These high-density populations are currently restricted to ~0.6 km2. Organisms living in small, isolated populations with very restricted ranges are at higher risk of extirpation due to various direct and indirect forces. Mark-recapture-resight surveys and distance sampling have been used to monitor the populations since 2010 and 2012 respectively. The data show that the high-density populations are declining. The current population size is estimated to be 4130-4860 individuals in 2015. A population viability analysis (PVA) was conducted to identify the most pressing threats and specific life history traits that are affecting this decline. The analysis estimates that if current trends persist, the species will be extinct in the wild in less than ten years. Adult mortality is a main factor and female mortality specifically characterizes this decline. In order for this species to persist over the next fifty years, adult mortality needs to be reduced by more than 50%. A lack of enforcement of the current laws results in the persistence of the main threat, poaching for consumption, thus altering the species distribution and causing high adult mortality. This is complicated by social customs and a lack of post primary education. Management changes could mitigate this threat and slow the population decline. Recommendations include an education campaign on the island, increased enforcement of the current laws, and breeding of C. oedirhina in situ and ex situ for release into the wild. / Includes bibliography. / Dissertation (Ph.D.)--Florida Atlantic University, 2015. / FAU Electronic Theses and Dissertations Collection