Natural Language Processing (NLP) develops computational approaches to processing language data. Supervised machine learning has become the dominant methodology of modern NLP. The performance of a supervised NLP system crucially depends on the amount of data available for training. In the standard supervised framework, if a sequence of words was not encountered in the training set, the system can only guess at its label at test time. The cost of producing labeled training examples is a bottleneck for current NLP technology. On the other hand, a vast quantity of unlabeled data is freely available. This dissertation proposes effective, efficient, versatile methodologies for 1) extracting useful information from very large (potentially web-scale) volumes of unlabeled data and 2) combining such information with standard supervised machine learning for NLP. We demonstrate novel ways to exploit unlabeled data, we scale these approaches to make use of all the text on the web, and we show improvements on a variety of challenging NLP tasks. This combination of learning from both labeled and unlabeled data is often referred to as semi-supervised learning. Although lacking manually-provided labels, the statistics of unlabeled patterns can often distinguish the correct label for an ambiguous test instance. In the first part of this dissertation, we propose to use the counts of unlabeled patterns as features in supervised classifiers, with these classifiers trained on varying amounts of labeled data. We propose a general approach for integrating information from multiple, overlapping sequences of context for lexical disambiguation problems. We also show how standard machine learning algorithms can be modified to incorporate a particular kind of prior knowledge: knowledge of effective weightings for count-based features. We also evaluate performance within and across domains for two generation and two analysis tasks, assessing the impact of combining web-scale counts with conventional features. In the second part of this dissertation, rather than using the aggregate statistics as features, we propose to use them to generate labeled training examples. By automatically labeling a large number of examples, we can train powerful discriminative models, leveraging fine-grained features of input words.

Wir versuchen herauszufinden, ob das subjektive Maß der Cloze-Vorhersagbarkeit mit der Kombination objektiver Maße (semantische und n-gram-Maße) geschätzt werden kann, die auf den statistischen Eigenschaften von Textkorpora beruhen. Die semantischen Maße werden entweder durch Abfragen von Internet-Suchmaschinen oder durch die Anwendung der Latent Semantic Analysis gebildet, während die n-gram-Wortmaße allein auf den Ergebnissen von Internet-Suchmaschinen basieren. Weiterhin untersuchen wir die Rolle der Cloze-Vorhersagbarkeit in SWIFT, einem Modell der Blickkontrolle, und wägen ab, ob andere Parameter den der Vorhersagbarkeit ersetzen können. Unsere Ergebnisse legen nahe, dass ein computationales Modell, welches Vorhersagbarkeitswerte berechnet, nicht nur Maße beachten muss, die die Relatiertheit eines Wortes zum Kontext darstellen; das Vorhandensein eines Maßes bezüglich der Nicht-Relatiertheit ist von ebenso großer Bedeutung. Obwohl hier jedoch nur Relatiertheits-Maße zur Verfügung stehen, sollte SWIFT ebensogute Ergebnisse liefern, wenn wir Cloze-Vorhersagbarkeit mit unseren Maßen ersetzen. / We try to determine whether it is possible to approximate the subjective Cloze predictability measure with two types of objective measures, semantic and word n-gram measures, based on the statistical properties of text corpora. The semantic measures are constructed either by querying Internet search engines or by applying Latent Semantic Analysis, while the word n-gram measures solely depend on the results of Internet search engines. We also analyse the role of Cloze predictability in the SWIFT eye movement model, and evaluate whether other parameters might be able to take the place of predictability. Our results suggest that a computational model that generates predictability values not only needs to use measures that can determine the relatedness of a word to its context; the presence of measures that assert unrelatedness is just as important. In spite of the fact, however, that we only have similarity measures, we predict that SWIFT should perform just as well when we replace Cloze predictability with our measures.

Background. Traditional plant use was studied in Lao PDR. Research focused on medicinal plant use by the Brou, Saek and Kry ethnic groups, traditional plant repellents against parasitic arthropods and leeches, and the phylogeny and biogeography of the medicinally-important snake gourd genus (Trichosanthes, Cucurbitaceae).  Methods. The ethnobiology research used a combination of structured interviews, village surveys, botanical collecting, hydro-distillation, GC-MS analysis, literature studies, and laboratory experiments. The plant systematics research used a combination of morphological studies, molecular biology laboratory work, and phylogenetic, dating and biogeographical analysis.  Results. Informants reported the use of close to 100 species to repel arthropods and leeches, many of which have constituents with documented efficacy.  Brou, Saek and Kry informants use over 75 plant species for women’s healthcare, mainly during the postpartum period for steam sauna, steam bath, hotbed, mother roasting, medicinal decoctions and infusions, and postpartum diet.  A molecular phylogeny of Trichosanthes and Gymnopetalum using a broad sampling of ~60% of their species and 4756 nucleotides of nuclear and plastid DNA shows that Gymnopetalum is nested within Trichosanthes. Fossil-calibrated Bayesian molecular dating of the Trichosanthes phylogeny reveals an early Oligocene origin of the genus, and many of the extant sections originating and diversifying during the Miocene. Biogeographical analysis shows a likely East or South Asian origin of Trichosanthes, with lineages diversifying and spreading throughout Australasia from the early Pliocene to the Pleistocene.  Discussion. Traditional plant use in Lao PDR is common and widespread. The presence among the repellent species of economical alternatives to costly synthetic repellents is tenable, and the subject of ongoing studies.  Postpartum traditions and medicinal plant use are essential parts of childbirth and postpartum recovery in these ethnic groups, and many other groups in Lao PDR. Efforts to improve maternal healthcare and reduce maternal and infant mortality need to integrate these traditions with modern notions of healthcare to achieve wider adoption. Documenting all possible uses of commonly used medicinal plant species shows that similarity in use between these ethnic groups is relatively low considering that they share, and have shared for many generations, the same environment and resources. A lack of effective cures leads to a process of continuous innovation, where effective cures are shared between cultures, but remedies of only cultural importance, or those under evaluation are culture-specific.  The Trichosanthes phylogeny implies the merging of Gymnopetalum into Trichosanthes, and this is done using available names or new combinations. A synopsis of Trichosanthes, the new combinations, and a revision of the species in Australia, are made and presented.  Conclusions. Traditional plant use is widespread in Lao PDR, and of significance to many people as a source of primary healthcare and inexpensive repellents. The important medicinal plant genus Trichosanthes includes Gymnopetalum, and has a complex biogeographic history with multiple colonization events of Australasia.

<p>This study examines how consumers perceive the overall quality and similarity/fit of the brand extensions (Nike camera, Nike socks, and Nike golf balls) in relation to the original brand (Nike athletic shoes) and how these perceptions influence their attitudes towards the extensions. The researcher proposes that the perceived overall quality of the brand extensions will be congruent to that of the original brand and that the attitude towards the brand extensions will be favourable only when there is a perceived similarity/fit between the brand extensions and the original brand. In order to get these insights, the study surveyed 147 undergraduate and postgraduate students from the University of the Western Cape across all faculties. A non-probability convenience sampling method was used to access respondents. To collect data, qualitative and quantitative methods were employed using a questionnaire which consisted of open-ended and closed-ended questions in the form of free associations technique and 5-point Likert scales. The data gathered was analysed by means of descriptive statistics and Spearman correlation coefficient. The findings indicate two things. (1) respondents only perceived Nike socks&rsquo / overall quality to be congruent to the original brand, Nike athletic shoes. (2) respondents only perceived Nike socks to be strongly similar to the original brand, Nike athletic shoes.</p>

Convection set up in a fluid due to the presence of two components of differing diffusivities is known as double diffusive convection. Double diffusive convection is observed in nature, in oceans, in the formation of certain columnar rock structures and in stellar interiors. The major engineering applications of double diffusive convection are in the fields metallurgy and alloy solidification in casting processes. The two components may be any two substances which affect the density of the fluid, heat and salt being the pair found most commonly in nature. Depending upon the initial stratifications of the two components, double diffusive convection can be set up in either the diffusive mode or the finger mode. In this thesis, the linear stability of a double diffusive system prone to finger instability has been studied in the presence of temporally varying non-linear background profiles of temperature and salinity. The motivation for the present study is to bridge the gap between existing theories, which mainly concentrate on linear background profiles independent of time, on the one hand and experiments and numerical simulations, which have time dependent step-like non-linear background profiles, on the other. The general stability characteristics of a double diffusive system with step-like background profiles have been studied using the standard normal mode method. The background temperature and salinity profiles are assumed to follow the hyperbolic tangent function, since it has a step-like character. The sharpness of the step can be altered by changing a suitable parameter in the hyperbolic tangent function. It is found that changing the degree of non-linearity of the background profile of one of the components keeping the background profile of the other component linear affects the growth rate, Wave number and the form of the disturbances. In general, increasing the degree of nonlinearity of background salinity profile makes the system more unstable and results in a reduction in the vertical extent of the disturbances. On the other hand, increasing the degree of non-linearity of the background temperature profile with the salinity profile kept linear results in a reduction in the growth rate and increase in the wave number. The form of the disturbance may change due to enhanced modal competition between the gravest odd and even modes in this case. The method of normal modes inherently assumes that the background profiles of temperature and salinity are independent of time and hence, it cannot be used for studying the stability of systems with time varying background profiles. A pseudo-similarity method has been used to handle such background profiles. Initial steps of temperature and salinity diffuse according to the error function form, and hence, the case of error function background profiles has been studied in detail. Taking into account the time-dependence of background profiles has been shown to significantly change the wave number and the incipient flux ratio. The dependence of the critical wave number (kc) on the thermal Rayleigh number (RaT ) can be determined analytically and is found to change from kc ~ Ra T1/4 for linear background profiles to kc ~ Ra T1/3 for error function profiles. The region of instability in the Rp (density stability ratio) space is found to increase from 1 ≤ R ρ ≤ r−1 for linear background profiles to 1 ≤ Rρ < r−3/2 for error function background profiles, where T denotes the ratio of the diffusivity of the slower diffusing component to that of the faster diffusing one. A parametric study covering a wide range of parameter values has been carried out to determine the effect of the parameters density stability ratio (Rp), diffusivity ratio (ρ ) and Prandtl number (Pr) on the onset time, critical wavenumber and the incipient flux ratio. The wide range of governing parameters covered here is beyond the scope of experimental and numerical studies. Such a wide range can be covered by theoretical approaches alone. It has been shown that the time of onset of convection determines the thicknesses of the temperature and salinity boundary layers, which in turn determine the width of salt fingers. Finally, the theoretical predictions of salt finger widths have been shown to be in agreement with the results of two dimensional numerical simulations of thermohaline system.

Il presente lavoro di ricerca si focalizza sul coping diadico, ovvero sulle modalità con cui partner affrontano come coppia, secondo la prospettiva di Guy Bodenmann (1997, 2000, 2005), le situazioni stressanti quotidiane. Primo obiettivo del presente lavoro di tesi è stato analizzare in un campione italiano (N = 778 partecipanti; cfr. Studio 1) la struttura fattoriale di uno strumento self-report messo a punto da Bodenmann (“Dyadic Coping Questionnaire”, Bodenmann, 1997, 2000) allo scopo di misurare la tendenza dei partner a mettere in atto diverse modalità di coping diadico. I risultati del primo studio confermano la struttura multifattoriale della scala, in linea con la teorizzazione di Bodenmann. In particolare si evince una più fine articolazione delle risposte di coping diadico positivo rispetto a quelle di coping diadico negativo. Nonostante l’importanza del coping per il benessere dei partner e della relazione (Bodenmann, Pihet, & Kaiser, 2006), poco si conosce dei possibili precursori di tale competenza relazionale. Secondo obiettivo del presente lavoro è stato dunque esaminare due delle possibili fonti dell’acquisizione del coping diadico da parte dei partner analizzando, in un campione composto da coppie in procinto di sposarsi e dai loro genitori (N = 764 partecipanti) se e quanto i partner fossero simili 1) ai loro rispettivi genitori e 2) tra loro per ciò che concerne la tendenza ad usare il coping diadico. Tali somiglianze sono inoltre state confrontate tra loro alla luce del processo di riallineamento caratteristico della fase della relazione che la coppia giovane sta attraversando (cfr. Studio 2). Due tipi di somiglianza sono stati presi in considerazione: la somiglianza unica e la somiglianza stereotipica, allo scopo di tenere in considerazione come il comune background culturale dei partecipanti possa influire sulle somiglianze esaminate. Somiglianza unica e stereotipica sono state qui considerate entrambe come portatrici di significato e sono state dunque esaminate parallelamente. Lo Studio 3 infine aveva l’ obiettivo di approfondire le somiglianze tra genitori e figli emerse nello studio precedente alla luce del genere del figlio, dal genere del genitore, e dalla valutazione da parte dei figli dei modelli di vita rappresentati dai loro genitori. I risultati hanno mostrato come sia le somiglianze tra figli e genitori sia tra i partner nella tendenza al coping diadico siano significative e come le somiglianze tra genitori e figli varino in funzione della dimensione di coping diadico considerata, del genere del figlio e del tipo di modello di coping diadico che i genitori rappresentano per i propri figli. Quanto emerso è stato inoltre discusso alla luce delle piste future di ricerca e implicazioni per l’intervento. / The present work is focused on dyadic coping, that is the way partners manage as a couple the stress they encounter in their everyday life (Bodenmann, 1997, 2000, 2005). The first objective of the present research was to examine in a sample of Italian couples (N = 778 participants) the factorial structure of a self-report instrument designed to measure partners’ dyadic coping tendency (Dyadic coping Questionnaire by Bodenmman, 1997, 2000). Findings from the first study of the present work confirmed the multidimensional nature of the scale, in line with Bodenmnann’s theory. In particular, it emerged a more detailed definition of positive dyadic coping responses, as compared to negative ones. Despite the importance of dyadic coping for the well-being of the relationship as well as of the partners themselves, little is known on how this competence originates. The second objective of the present research was then to explore two possible sources of dyadic coping acquisition by examining whether and how young adults prior to marriage were similar to 1) their parents and 2) their partners in their dyadic coping tendency (N = 764 participants). Moreover, similarities with parents and with partners were compared in light of the specific stage of the relationship partners were living (cfr. Study 2). Similarity was computed adopting an idiographic approach via intraclass correlations and stereotype adjustment was performed in order to take into account the impact of partners and parents shared cultural background on similarity. For the purpose of the present work both stereotypical and unique similarities were considered meaningful and then examined simultaneously. Study 3 was aimed at further exploring the parent-child similarities emerged from study 2 by assessing whether and how parent-child similarity in dyadic coping differed as a function of parents and children’s gender and whether they were associated with children’s perceptions of the kind of models their parents represented for their lives. Results highlight that both parent-child and partners’ similarities were significant and differently modulated as a function of the different dimensions of dyadic coping considered, children’s gender as well as the kind of dyadic coping models parents represented for their children. Future paths of research and implications for intervention were discussed.

<p>Nearly one-third of the world&rsquo / s tortoises live in South Africa, but little is known about their habitat requirements and feeding ecology. Chersina angulata, the angulate tortoise, is endemic to&nbsp / southern Africa, with a wide distribution along the western and southern coasts. Because this tortoise occupies a number of different habitat types, it has always been considered a generalist&nbsp / herbivore, although little is known about its&nbsp / diet and other needs. This study evaluates the habitat characteristics and feeding ecology of C. angulata at two study sites in the southwestern&nbsp / &nbsp / &nbsp / Cape, the West Coast National Park (WCNP) and Dassen Island (DI). The WCNP is a large conserved area in the Fynbos biome, along the southwestern coast of South Africa, whereas DI is a&nbsp / small offshore island with low floral and faunal diversity, just south of the WCNP. The efficacy of three methods used to study the feeding ecology of herbivores, focal observations, macroscopic faecal analysis and histological analysis of scats, was evaluated. Plant cover, species diversity, and the variety of growth forms were substantially larger at the WCNP than on DI.&nbsp / In the WCNP, shrubs and grasses were the dominant growth forms but the vegetation also included herbs, succulents, restios, sedges and parasitic plants. A few perennial species such as&nbsp / the grass Ehrharta villosa, shrubs such as Helichrysum niveum, Nylandtia spinosa and Rhus spp., and succulents such as Carpobrotus edulis and Ruschia spp., provided most of the plant&nbsp / cover. DI had a depauperate flora, consisting of succulents and herbs, and ephemeral plants contributed more than perennials did to plant cover throughout the year. The succulents Mesembryanthemum crystallinum and Tetragonia fruticosa provided most of the cover on DI. Angulate tortoises are herbivores and 72 diet plants in 32 plant families were identified to the&nbsp / species or genus level. Several diet species, however, could not be identified. In&nbsp / addition to angiosperms, the tortoises&rsquo / diet included mosses, mushrooms, insects,snails and animal faeces.&nbsp / The most important growth forms in the diet were herbs and grasses. The diet of the WCNP tortoises was more diverse than the diet of DI tortoises, but the number of principal food items in&nbsp / the diet did not differ between the two sites. Over an annual cycle, WCNP tortoises had four principal food plants while DI tortoises had five principal food plants. At both sites, principal food&nbsp / &nbsp / plants changed with the season and few plants remained principal food items in more than one season. Cynodon dactylon was a principal food item in three of the four seasons in the WCNP, whereas Trachyandra divaricata was a principal food plant each season on DI. Most principal food plants were grass or herb species but the sedge Ficinia nigrescens, and a succulent that&nbsp / could be identified only to the family level (Aizoaceae), featured strongly in the spring diets of DI and WCNP tortoises, respectively.&nbsp / The three study methods did not provide the same type or quality of information about the feeding ecology of angulate tortoises. The small size and wary nature of angulate tortoises compromised focal studies because it was often not possible to see&nbsp / what the tortoises ate. This method, however, provided the interesting observation that rabbit&nbsp / faecal pellets contributed nearly 30% to summer and autumn diets on DI when food was scarce.&nbsp / Rabbit faeces may not only provide a source of nutrients but may also supplement the microflora, required to digest cellulose, in the tortoises&rsquo / guts. Macroscopic evaluation of the tortoises&rsquo / &nbsp / scats appeared to be an ineffective method to identify diet plants, and the bulk of the scat mass could not be identified. This indicates&nbsp / that angulate tortoises either selected food low in fibrous&nbsp / content or that the digestive system of the tortoises dealt efficiently with tough plant material. The macroscopic method was the only method that highlighted the large contribution of&nbsp / fruits / seeds to the diet of angulate tortoises. Since the tortoises digested many seeds only partially, or not at all, C. angulata is potentially an important agent of seed dispersal in the southwestern Cape. The macroscopic study showed that on DI, sand made up 28% of the scat mass in spring, whereas sand never made a substantial contribution to the scat composition of WCNP tortoises. Lithophagy may be an important strategy in a depauperate habitat, such as DI, because the abrasive action of sand may help with the digestion of tough plants, or the sand may&nbsp / provide the tortoises with important minerals that are deficient in their food plants.The histological analysis of scats provided the most comprehensive diet list for C. angulata. Selection indices&nbsp / based on data from the histological analysis indicated that angulate tortoises were highly selective in their food choice. Most of the principal food items were selected out of proportion to their&nbsp / availability and the tortoises avoided the most abundant plants in their habitats. Several factors, such as palatability, accessibility and profitability, may have influenced their food choice. The proportional similarity indices for WCNP and DI tortoises, respectively, were 0.31 and 0.16, confirming that C. angulata is a food specialist and not a food generalist as was previously thought. This factor should be considered in the management of this species and in future conservation planning of its habitat.&nbsp / &nbsp / </p>

Traditional multiple hypothesis testing procedures, such as that of Benjamini and Hochberg, fix an error rate and determine the corresponding rejection region. In 2002 Storey proposed a fixed rejection region procedure and showed numerically that it can gain more power than the fixed error rate procedure of Benjamini and Hochberg while controlling the same false discovery rate (FDR). In this thesis it is proved that when the number of alternatives is small compared to the total number of hypotheses, Storeys method can be less powerful than that of Benjamini and Hochberg. Moreover, the two procedures are compared by setting them to produce the same FDR. The difference in power between Storeys procedure and that of Benjamini and Hochberg is near zero when the distance between the null and alternative distributions is large, but Benjamini and Hochbergs procedure becomes more powerful as the distance decreases. It is shown that modifying the Benjamini and Hochberg procedure to incorporate an estimate of the proportion of true null hypotheses as proposed by Black gives a procedure with superior power.<p> Multiple hypothesis testing can also be applied to regression diagnostics. In this thesis, a Bayesian method is proposed to test multiple hypotheses, of which the i-th null and alternative hypotheses are that the i-th observation is not an outlier versus it is, for i=1,...,m. In the proposed Bayesian model, it is assumed that outliers have a mean shift, where the proportion of outliers and the mean shift respectively follow a Beta prior distribution and a normal prior distribution. It is proved in the thesis that for the proposed model, when there exists more than one outlier, the marginal distributions of the deletion residual of the i-th observation under both null and alternative hypotheses are doubly noncentral t distributions. The outlyingness of the i-th observation is measured by the marginal posterior probability that the i-th observation is an outlier given its deletion residual. An importance sampling method is proposed to calculate this probability. This method requires the computation of the density of the doubly noncentral F distribution and this is approximated using Patnaiks approximation. An algorithm is proposed in this thesis to examine the accuracy of Patnaiks approximation. The comparison of this algorithms output with Patnaiks approximation shows that the latter can save massive computation time without losing much accuracy.<p> The proposed Bayesian multiple outlier identification procedure is applied to some simulated data sets. Various simulation and prior parameters are used to study the sensitivity of the posteriors to the priors. The area under the ROC curves (AUC) is calculated for each combination of parameters. A factorial design analysis on AUC is carried out by choosing various simulation and prior parameters as factors. The resulting AUC values are high for various selected parameters, indicating that the proposed method can identify the majority of outliers within tolerable errors. The results of the factorial design show that the priors do not have much effect on the marginal posterior probability as long as the sample size is not too small.<p> In this thesis, the proposed Bayesian procedure is also applied to a real data set obtained by Kanduc et al. in 2008. The proteomes of thirty viruses examined by Kanduc et al. are found to share a high number of pentapeptide overlaps to the human proteome. In a linear regression analysis of the level of viral overlaps to the human proteome and the length of viral proteome, it is reported by Kanduc et al. that among the thirty viruses, human T-lymphotropic virus 1, Rubella virus, and hepatitis C virus, present relatively higher levels of overlaps with the human proteome than the predicted level of overlaps. The results obtained using the proposed procedure indicate that the four viruses with extremely large sizes (Human herpesvirus 4, Human herpesvirus 6, Variola virus, and Human herpesvirus 5) are more likely to be the outliers than the three reported viruses. The results with thefour extreme viruses deleted confirm the claim of Kanduc et al.

Systems can become misconfigured for a variety of reasons such as operator errors or buggy patches. When a misconfiguration is discovered, usually the first order of business is to restore availability, often by undoing the misconfiguration. To simplify this task, we propose Ocasta to automatically determine which files contain configuration state. Ocasta uses a novel {\em similarity} metric to measures how similar a file's versions are to each other, and a set of filters to eliminate non-persistent files from consideration. These two mechanisms enable Ocasta to identify all 72 configuration files out of 2363 versioned files from 6 common applications in two user traces, while mistaking only 33 non-configuration files as configuration files. Ocasta allows a versioning file system to eliminate roughly 66\% of non-configuration file versions from its logs, thus reducing the number of file versions that a user must manually examine to recover from a misconfiguration.

Heuristic Evaluation (HE) is a popular usability inspection method. Yet little is known about the effect the evaluators have on the outcome of HE. One potentially important feature of evaluators is their end-user status, that is, whether or not they are end-users for whom the interface is designed. I completed a detailed review of the HE literature, combined sources, developed an explicit method for conducting an HE and trained HE novices from different work domains using it. Using these methods I conducted a preliminary randomized crossover study (n=6) of the effect of end-user status during the inspection and merging stages of HE. I estimate a larger study of approximately 148 end-users would be needed to test hypotheses regarding end-user status. I demonstrated a novel measure of the effect of end-user status for the merging stage of HE, which I called the measure of matching similarity (MMS).