Spelling suggestions: "subject:"categorical"" "subject:"categoricals""
241 |
Node Classification on Relational Graphs Using Deep-RGCNsChandra, Nagasai 01 March 2021 (has links) (PDF)
Knowledge Graphs are fascinating concepts in machine learning as they can hold usefully structured information in the form of entities and their relations. Despite the valuable applications of such graphs, most knowledge bases remain incomplete. This missing information harms downstream applications such as information retrieval and opens a window for research in statistical relational learning tasks such as node classification and link prediction. This work proposes a deep learning framework based on existing relational convolutional (R-GCN) layers to learn on highly multi-relational data characteristic of realistic knowledge graphs for node property classification tasks. We propose a deep and improved variant, Deep-RGCNs, with dense and residual skip connections between layers. These skip connections are known to be very successful with popular deep CNN-architectures such as ResNet and DenseNet. In our experiments, we investigate and compare the performance of Deep-RGCN with different baselines on multi-relational graph benchmark datasets, AIFB and MUTAG, and show how the deep architecture boosts the performance in the task of node property classification. We also study the training performance of Deep-RGCNs (with N layers) and discuss the gradient vanishing and over-smoothing problems common to deeper GCN architectures.
|
242 |
Regression Analysis for Ordinal Outcomes in Matched Study Design: Applications to Alzheimer's Disease StudiesAustin, Elizabeth 09 July 2018 (has links) (PDF)
Alzheimer's Disease (AD) affects nearly 5.4 million Americans as of 2016 and is the most common form of dementia. The disease is characterized by the presence of neurofibrillary tangles and amyloid plaques [1]. The amount of plaques are measured by Braak stage, post-mortem. It is known that AD is positively associated with hypercholesterolemia [16]. As statins are the most widely used cholesterol-lowering drug, there may be associations between statin use and AD. We hypothesize that those who use statins, specifically lipophilic statins, are more likely to have a low Braak stage in post-mortem analysis.
In order to address this hypothesis, we wished to fit a regression model for ordinal outcomes (e.g., high, moderate, or low Braak stage) using data collected from the National Alzheimer's Coordinating Center (NACC) autopsy cohort. As the outcomes were matched on the length of follow-up, a conditional likelihood-based method is often used to estimate the regression coefficients. However, it can be challenging to solve the conditional-likelihood based estimating equation numerically, especially when there are many matching strata. Given that the likelihood of a conditional logistic regression model is equivalent to the partial likelihood from a stratified Cox proportional hazard model, the existing R function for a Cox model, coxph( ), can be used for estimation of a conditional logistic regression model. We would like to investigate whether this strategy could be extended to a regression model for ordinal outcomes.
More specifically, our aims are to (1) demonstrate the equivalence between the exact partial likelihood of a stratified discrete time Cox proportional hazards model and the likelihood of a conditional logistic regression model, (2) prove equivalence, or lack there-of, between the exact partial likelihood of a stratified discrete time Cox proportional hazards model and the conditional likelihood of models appropriate for multiple ordinal outcomes: an adjacent categories model, a continuation-ratio model, and a cumulative logit model, and (3) clarify how to set up stratified discrete time Cox proportional hazards model for multiple ordinal outcomes with matching using the existing coxph( ) R function and interpret the regression coefficient estimates that result. We verified this theoretical proof through simulation studies. We simulated data from the three models of interest: an adjacent categories model, a continuation-ratio model, and a cumulative logit model. We fit a Cox model using the existing coxph( ) R function to the simulated data produced by each model. We then compared the coefficient estimates obtained. Lastly, we fit a Cox model to the NACC dataset. We used Braak stage as the outcome variables, having three ordinal categories. We included predictors for age at death, sex, genotype, education, comorbidities, number of days having taken lipophilic statins, number of days having taken hydrophilic statins, and time to death. We matched cases to controls on the length of follow up. We have discussed all findings and their implications in detail.
|
243 |
Contributions to fuzzy object comparison and applications. Similarity measures for fuzzy and heterogeneous data and their applications.Bashon, Yasmina M. January 2013 (has links)
This thesis makes an original contribution to knowledge in the fi eld
of data objects' comparison where the objects are described by attributes
of fuzzy or heterogeneous (numeric and symbolic) data types.
Many real world database systems and applications require information
management components that provide support for managing
such imperfect and heterogeneous data objects. For example,
with new online information made available from various sources, in
semi-structured, structured or unstructured representations, new information
usage and search algorithms must consider where such data
collections may contain objects/records with di fferent types of data:
fuzzy, numerical and categorical for the same attributes.
New approaches of similarity have been presented in this research to
support such data comparison. A generalisation of both geometric and set theoretical similarity models has enabled propose new similarity
measures presented in this thesis, to handle the vagueness (fuzzy data
type) within data objects. A framework of new and unif ied similarity
measures for comparing heterogeneous objects described by numerical,
categorical and fuzzy attributes has also been introduced.
Examples are used to illustrate, compare and discuss the applications
and e fficiency of the proposed approaches to heterogeneous data comparison. / Libyan Embassy
|
244 |
Exploring Hybridity in the 21st Century: The Working Lives of South Asian Ethnic Minorities from a British Born Generation in Bradford.Rifet, Saima January 2015 (has links)
This thesis explores the working lives of British Born South Asian Ethnic
Minorities (BB SAEMs), critiquing the homogenous identities ascribed to them in
previous research. Its methodology is life-story interviews analysed using Nvivo.
This identified four hybrid categories emerging from two cultures. I fitted myself
neatly into just one. However the reflexive analysis required in good qualitative
research led me to realise that I fitted into not one, but all four categories, and
into others not yet recognised. At this point, my thesis had to take a new turn.
An auto-ethnographic, moment-by-moment study led to an ‘unhybrid
categorisation of hybridities’ acknowledging ‘fuzziness and mélange, cut ‘n’ mix,
and criss and crossover’ where identity is a complex-mix, always in flux. I
conclude not only with this new theory of identity formation in the working lives
of BB SAEMs, but also by arguing that by imposing the requirement to
categorise, research methods lead to over-simplification and misunderstanding. / University of Bradford
|
245 |
Сайт православного прихода как сверхтекст : магистерская диссертация / Orthodox parish website as supertextБердникова, В. А., Berdnikova, V. A. January 2023 (has links)
Работа выполнена в русле развития функциональной стилистики. В исследовании анализируются тексты, расположенные на сайтах православных приходов. В диссертации применяется категориально-текстовой метод, разработанный в рамках школы лингвокультурологии и стилистики УрФУ Т. В. Матвеевой. Развивается идея о том, что религиозный текст на сайте является единицей сверхтекста. Для анализа выбраны категории темы, локальности и темпоральности. В ходе исследования автор заключает, что рассматриваемые тексты обладают содержательной целостностью, ограничены локально и темпорально, что соотносится с признаками сверхтекста. Проведенный анализ позволяет сделать вывод о том, что тексты на официальных сайтах приходов являются частями сверхтекста, подчиненного конструктивному принципу прототекстуальности и сохраняющему в той или иной степени черты религиозного стиля. / The work is carried out in line with the development of functional stylistics. The study analyzes the texts located on the websites of Orthodox parishes. The thesis applies the categorical-textual method developed within the school of linguoculturalism and stylistics of UrFU by T. V. Matveeva. The idea that the religious text on the website is a unit of supertext is developed. The categories of theme, locality and temporality are chosen for the analysis. In the course of the study the author concludes that the texts in question have a substantive integrity, limited locally and temporally, which correlates with the signs of the supertext. The analysis carried out allows us to conclude that the texts on the official websites of parishes are parts of the supertext, subordinate to the constructive principle of prototextuality and preserving the features of religious style to a greater or lesser extent.
|
246 |
Latent Variable Models of Categorical Responses in the Bayesian and Frequentist FrameworksFarouni, Tarek January 2014 (has links)
No description available.
|
247 |
Connecting Unsupervised and Supervised Categorization Behavior from a Parainformative PerspectiveDoan, Charles A. 12 June 2018 (has links)
No description available.
|
248 |
Visualization of Clustering Solutions for Large Multi-dimensional Sequential DatasetsDornala, Maninder 29 May 2018 (has links)
No description available.
|
249 |
Syntax, semantics, and pragmatics of accusative-quotative constructions in JapaneseHorn, Stephen Wright 19 March 2008 (has links)
No description available.
|
250 |
Fair Partitioning of Procedurally Generated Game Maps for Grand Strategy GamesOttander, Jens January 2022 (has links)
Due to the high cost of manual content creation within the game development industry, methods for procedural generation of content such as game maps and levels have emerged. However, methods for generating game maps have remained relatively unexplored in competitive multiplayer contexts. Presumably, this is due to the opposing goals of generating game maps that are both interesting and fair. This study aims to explore the possibility of satisfying both these goals simultaneously by separating the generative phase from the phase that enforces fairness. In this endeavor, simple game maps for a generic multiplayer grand strategy game are generated using noise-based methods. The task of partitioning the game map fairly between the players is then modeled as a constrained categorical multiobjective minimization problem that is subsequently solved by two genetic algorithms, the reference-point-based algorithm NSGA-III and the decomposition-based algorithm MOEA/D-IEpsilon. In a primary study, the proposed partitioning method is evaluated based on the quality of the solutions produced, its scalability, and its ability to find symmetrical partitions of symmetrical game maps. The results show that the proposed method makes significant improvement from the initial guess but fails to produce completely fair partitions in general. Explanations and possible solutions to this are presented. The timing results indicate that the proposed method is not applicable in real-time contexts. However, the proposed method might still be applicable in online contexts if smaller game maps are considered and in offline contexts if larger game maps are considered. Finally, the partitioning results show that the proposed method successfully finds fair partitions of symmetrical game maps but fails to find the obvious symmetrical partitions. In a secondary study, the two genetic algorithms are compared to determine which algorithm produces dominating solutions and which algorithm produces most diverse solution. The results indicate that, for the partitioning problems considered in this study, the reference-point-based algorithm is both dominant and produces the most diverse solutions.
|
Page generated in 0.043 seconds