No description available.

No description available.

A new methodology is introduced where the natural hierarchical character of model descriptions and simulation results are exploited to expedite analysis of problems. The philosophy and the different concepts involved are illustrated by implementing the strategy to solve some practical problems. The end result was a mix of mechanics, well-designed data structures and software interfaces that forms a rapid analysis environment. This can be very advantageous for cases where a sequence of analyses is required because of safety concerns or cost. When designing a structure, it is common to make frequent modifications to the model during the process. In such cases, the ability to use data from different models within the same analysis environment becomes a major advantage. The proposed system's forte is its hierarchical framework that allows models to communicate with each other and share information with one another. This makes it ideal for global local analyses where solutions from a global model are used to derive the boundary conditions for the local model. The system was also used to conduct a micro mechanical analysis on unidirectional composites that have a non-uniform spatial distribution of the fibers. The hierarchical strategy is not tied to any specific methodology and can be adapted to solve problem using different technologies. This allows the strategy to be used across multiple length scales and governing equations.

A new methodology is introduced where the natural hierarchical character of model descriptions and simulation results are exploited to expedite analysis of problems. The philosophy and the different concepts involved are illustrated by implementing the strategy to solve some practical problems. The end result was a mix of mechanics, well-designed data structures and software interfaces that forms a rapid analysis environment. This can be very advantageous for cases where a sequence of analyses is required because of safety concerns or cost. When designing a structure, it is common to make frequent modifications to the model during the process. In such cases, the ability to use data from different models within the same analysis environment becomes a major advantage. The proposed system's forte is its hierarchical framework that allows models to communicate with each other and share information with one another. This makes it ideal for global local analyses where solutions from a global model are used to derive the boundary conditions for the local model. The system was also used to conduct a micro mechanical analysis on unidirectional composites that have a non-uniform spatial distribution of the fibers. The hierarchical strategy is not tied to any specific methodology and can be adapted to solve problem using different technologies. This allows the strategy to be used across multiple length scales and governing equations.

Clustering microarray data is a helpful way of identifying genes which are biologically related. Unfortunately, when attempting to cluster microarray data, certain issues must be considered including: the uncertainty in the number of true clusters; the expression of a given gene is often a ected by the expression of other genes; and microarray data is usually high dimensional. This thesis outlines a Bayesian in nite Gaussian mixture model which addresses the issues outlined above by: not requiring the researcher to specify the number of clusters expected, applying a non-diagonal covariance structure, and using mixtures of factor analyzers and extensions thereof to structure the covariance matrix such that it is based on a few latent variables. This approach will be illustrated on real and simulated data.

No description available.

No description available.

Many adult concepts can be represented in taxonomies hierarchical systems in which concepts are differentiated into varying levels of abstraction (e.g., musical instrument, wind instrument, flute) related by class inclusion (a flute is a wind instrument and a wind instrument is a musical instrument). Indeed, most natural kinds (e.g., whale, tree) and artifacts (e.g., flute, truck) are generally believed to fall within taxonomies. Moreover, in real world contexts, concepts are probably rarely learned as explicitly contrasting sets existing completely outside of known taxonomies (that is, one might not learn cats vs. dogs without also learning that both are types of animals, and that both include more specific subcategories). Surprisingly, relatively little research has been done on the learning of categories that are hierarchically structured. The present study began an investigation into how adults learned new concepts that are hierarchically structured. In Experiment 1, participants learned to classify items at one taxonomic level then at a later time classified items at either the same or a different level. The results suggested that people were unable to clearly detect the relationship among alternate levels of the hierarchy prior to exposure of those levels. However, results in Experiment 1 also suggested that learning multiple categories might lead to deeper understanding of how features transfer or generalize to higher taxonomic levels. The remaining experiments addressed more explicitly the influence of hierarchical structures on category learning by including prototype and control items, along with artificial and knowledge-based category labels. Results from these experiments indicated that, at least within the parameters of this study, prior experience cued by knowledge-based category labels interacted adversely with abstract materials and interfered with mapping of item information to categories. Moreover, when the relationship between categories and item information is unclear, generalization might be one important means by which people categorize.

Master of Science / Department of Computing and Information Sciences / William H. Hsu / Clustering in data mining is a process of discovering groups in a set of data such that the similarity within the group is maximized and the similarity among the groups is minimized. One way of approaching clustering is to treat it as a blocking problem of minimizing the maximum distance between any two units within the same group. This method is known as Threshold blocking. It works by applying blocking as a graph partition problem. Chameleon is a hierarchical clustering algorithm, that based on dynamic modelling measures the similarity between two clusters. In the clustering process, to merge two cluster, we check if the inter-connectivity and closeness between two clusters are high relative to the internal inter-connectivity of the clusters and closeness of items within the clusters. This way of merging of cluster using the dynamic model helps in discovery of natural and homogeneous clusters. The main goal of this project is to implement a local implementation of CHAMELEON and compare the output generated from Chameleon against Threshold blocking algorithm suggested by Higgins et al with its hybridized form and unhybridized form.

Grizzly bear (Ursus arctos) den-site selection may be influenced by multiple factors at varying spatial extents. To test for hierarchical selection, I compared grizzly bear den sites (n = 21) to random locations in the Southern Lakes, Yukon at two scales: mesoscale (within 1 km of dens) and microscale (within 100 m of dens). I observed selection for den opening slope, pixel slope, soil content, and horizontal cover at the mesoscale. At the microscale, I observed selection for opening slope and horizontal cover. To further test the role of security and the sex hypothesis of sexual segregation, I compared adult female (n = 142), adult male (n = 36), and juvenile (n = 35) den locations in Denali National Park and Preserve, Alaska. Adult females spatially segregated from adult males, with dens characterized by higher elevations and steeper slopes than adult male den sites.