Matching Items (43,913)
Description
As the information available to lay users through autonomous data sources continues to increase, mediators become important to ensure that the wealth of information available is tapped effectively. A key challenge that these information mediators need to handle is the varying levels of incompleteness in the underlying databases in terms of missing attribute values. Existing approaches such as Query Processing over Incomplete Autonomous Databases (QPIAD) aim to mine and use Approximate Functional Dependencies (AFDs) to predict and retrieve relevant incomplete tuples. These approaches make independence assumptions about missing values--which critically hobbles their performance when there are tuples containing missing values for multiple correlated attributes. In this thesis, I present a principled probabilis- tic alternative that views an incomplete tuple as defining a distribution over the complete tuples that it stands for. I learn this distribution in terms of Bayes networks. My approach involves min- ing/"learning" Bayes networks from a sample of the database, and using it do both imputation (predict a missing value) and query rewriting (retrieve relevant results with incompleteness on the query-constrained attributes, when the data sources are autonomous). I present empirical studies to demonstrate that (i) at higher levels of incompleteness, when multiple attribute values are missing, Bayes networks do provide a significantly higher classification accuracy and (ii) the relevant possible answers retrieved by the queries reformulated using Bayes networks provide higher precision and recall than AFDs while keeping query processing costs manageable.
ContributorsRaghunathan, Rohit (Author) / Kambhampati, Subbarao (Thesis advisor) / Liu, Huan (Committee member) / Lee, Joohyung (Committee member) / Arizona State University (Publisher)
Created2011
Description
The San Gabriel Mountains (SGM) of southern California provide the opportunity to study the topographic controls on erosion rate in a mountain range where climate and lithology are relatively constant. I use a combination of digital elevation model data, detailed channel survey data, decadal climate records, and catchment-averaged erosion rates quantified from 10Be concentrations in stream sands to investigate the style and rates of hillslope and channel processes across the transition from soil-mantled to rocky landscapes in the SGM. Specifically, I investigate (1) the interrelations among different topographic metrics and their variation with erosion rate, (2) how hillslopes respond to tectonic forcing in "threshold" landscapes, (3) the role of discharge variability and erosion thresholds in controlling the relationship between relief and erosion rate, and (4) the style and pace of transient adjustment in the western SGM to a recent increase in uplift rate. Millennial erosion rates in the SGM range from 0.03-1.1 mm/a, generally increasing from west to east. For low erosion rates (< 0.3 mm/a), hillslopes tend to be soil-mantled, and catchment-averaged erosion rates are positively correlated with catchment-averaged slope, channel steepness, and local relief. For erosion rates greater than 0.3 mm/a, hillslopes become increasingly rocky, catchment-mean hillslope angle becomes much less sensitive to erosion rate, and channels continue to steepen. I find that a non-linear relationship observed between channel steepness and erosion rate can be explained by a simple bedrock incision model that combines a threshold for erosion with a probability distribution of discharge events where large floods follow an inverse power-law. I also find that the timing of a two-staged increase in uplift rate in the western SGM based on stream profile analysis agrees with independent estimates. Field observations in the same region suggest that the relict topography that allows for this calculation has persisted for more than 7 Ma due to the stalling of migrating knickpoints by locally stronger bedrock and a lack of coarse sediment cover.
ContributorsDibiase, Roman Alexander (Author) / Whipple, Kelin X (Thesis advisor) / Heimsath, Arjun M. (Thesis advisor) / Arrowsmith, J Ramon (Committee member) / Garnero, Edward J. (Committee member) / Hodges, Kip V. (Committee member) / Arizona State University (Publisher)
Created2011
Description
The repression of reproductive competition and the enforcement of altruism are key components to the success of animal societies. Eusocial insects are defined by having a reproductive division of labor, in which reproduction is relegated to one or few individuals while the rest of the group members maintain the colony and help raise offspring. However, workers have retained the ability to reproduce in most insect societies. In the social Hymenoptera, due to haplodiploidy, workers can lay unfertilized male destined eggs without mating. Potential conflict between workers and queens can arise over male production, and policing behaviors performed by nestmate workers and queens are a means of repressing worker reproduction. This work describes the means and results of the regulation of worker reproduction in the ant species Aphaenogaster cockerelli. Through manipulative laboratory studies on mature colonies, the lack of egg policing and the presence of physical policing by both workers and queens of this species are described. Through chemical analysis and artificial chemical treatments, the role of cuticular hydrocarbons as indicators of fertility status and the informational basis of policing in this species is demonstrated. An additional queen-specific chemical signal in the Dufour's gland is discovered to be used to direct nestmate aggression towards reproductive competitors. Finally, the level of actual worker-derived males in field colonies is measured. Together, these studies demonstrate the effectiveness of policing behaviors on the suppression of worker reproduction in a social insect species, and provide an example of how punishment and the threat of punishment is a powerful force in maintaining cooperative societies.
ContributorsSmith, Adrian A. (Author) / Liebig, Juergen (Thesis advisor) / Hoelldobler, Bert (Thesis advisor) / Gadau, Juergen (Committee member) / Johnson, Robert A. (Committee member) / Pratt, Stephen (Committee member) / Arizona State University (Publisher)
Created2011
Description
Chronic illness can affect multiple domains of functioning, yet scientific understanding of the effects across the lifespan and under multiple contexts is still developing. For instance, research consistently indicates the early years of a child's life are pivotal for early intervening to positively affect physical, cognitive, and socio-emotional development; unfortunately, the impact of chronic illnesses, and thus appropriate interventions, during this time are not well-established. Academic achievement is one area in which children with chronic illness are negatively affected and research suggests that the effects of illness can be exacerbated by certain social determinants of health and demographic characteristics; however, no recent studies have examined these relationships for children at school entry. The current study utilized the Early Childhood Longitudinal Study-Birth Cohort (ECLS-B) to examine variations in early academic readiness in reading and mathematics by diabetes status, race, and social determinants, specifically mother's education and access to early childhood education, among children born in 2001. Results of the current study indicated that children with diabetes scored lower on reading and mathematics relative to their non-diabetic peers. Significant interactions were evident for diabetes status by mother's education, race/ethnicity, and by early childhood education. Children in homes whose mothers had the lowest level of education did not score as high as children in homes with mothers who had higher levels of education. Among children without diabetes, those identified as Asian, Pacific Islander, or Native Hawaiian outperformed White, Black, Hispanic, American Indian, and multi-race groups on measures of reading and mathematics, whereas among children with diabetes, those identified as multiracial scored highest. Regardless of diabetes status, children who attended preschool outperformed those who did not, yet children without diabetes who had not attended preschool outperformed diabetic children who did receive such services. Findings support the need for targeted early intervention as preschool alone did not mitigate the effects of diabetes on academic performance.
ContributorsKucera, Miranda (Author) / Sullivan, Amanda L (Thesis advisor) / Wodrich, David (Committee member) / Thompson, Marilyn (Committee member) / Arizona State University (Publisher)
Created2012
Description
Higher education institutions in the state of Arizona have experienced a reduction in government funding due to the economic challenges the state is facing combined with an ongoing national recession. Three higher education institutions studied are located in Phoenix, Arizona. The three higher education institutions are Phoenix College, Arizona State University and The University of Phoenix. An analysis of documents made public by each institution was conducted and high level administrators at each institution were interviewed to learn about revenue streams currently active and planned. The results of this set of analyses were presented to the leadership team of Phoenix College in a three-hour strategic planning and priority setting meeting. The action research study assisted Phoenix College administrators in gaining knowledge about and initiating action plans to increase revenue through entrepreneurial strategies. Increased funding is necessary to offset reductions in state aid and property tax revenues. Implementing entrepreneurial strategies to increase funding can promote self-reliance and flexibility and can mitigate the damage to institutional mission success that is threatened by reductions in traditional funding. The strategic planning and priority setting exercise conducted at Phoenix College produced three immediate outcomes: it informed the community of practice about entrepreneurial strategies to increase funding that are in use by higher education institutions located in greater Phoenix, Arizona; it influenced the community of practice to examine entrepreneurial revenue streams and; it committed the leadership team to pursuing and enlarging three additional revenue streams.
ContributorsKakar, Casandra Gwen (Author) / Clark, Christopher M. (Thesis advisor) / De Los Santos Jr., Alfredo G. (Committee member) / Giovannini, Eugene (Committee member) / Arizona State University (Publisher)
Created2011
Description
In this thesis I introduce a new direction to computing using nonlinear chaotic dynamics. The main idea is rich dynamics of a chaotic system enables us to (1) build better computers that have a flexible instruction set, and (2) carry out computation that conventional computers are not good at it. Here I start from the theory, explaining how one can build a computing logic block using a chaotic system, and then I introduce a new theoretical analysis for chaos computing. Specifically, I demonstrate how unstable periodic orbits and a model based on them explains and predicts how and how well a chaotic system can do computation. Furthermore, since unstable periodic orbits and their stability measures in terms of eigenvalues are extractable from experimental times series, I develop a time series technique for modeling and predicting chaos computing from a given time series of a chaotic system. After building a theoretical framework for chaos computing I proceed to architecture of these chaos-computing blocks to build a sophisticated computing system out of them. I describe how one can arrange and organize these chaos-based blocks to build a computer. I propose a brand new computer architecture using chaos computing, which shifts the limits of conventional computers by introducing flexible instruction set. Our new chaos based computer has a flexible instruction set, meaning that the user can load its desired instruction set to the computer to reconfigure the computer to be an implementation for the desired instruction set. Apart from direct application of chaos theory in generic computation, the application of chaos theory to speech processing is explained and a novel application for chaos theory in speech coding and synthesizing is introduced. More specifically it is demonstrated how a chaotic system can model the natural turbulent flow of the air in the human speech production system and how chaotic orbits can be used to excite a vocal tract model. Also as another approach to build computing system based on nonlinear system, the idea of Logical Stochastic Resonance is studied and adapted to an autoregulatory gene network in the bacteriophage λ.
ContributorsKia, Behnam (Author) / Ditto, William (Thesis advisor) / Huang, Liang (Committee member) / Lai, Ying-Cheng (Committee member) / Helms Tillery, Stephen (Committee member) / Arizona State University (Publisher)
Created2011
Description
Raman scattering from Ge-Si core-shell nanowires is investigated theoretically and experimentally. A theoretical model that makes it possible to extract quantitative strain information from the measured Raman spectra is presented for the first time. Geometrical and elastic simplifications are introduced to keep the model analytical, which facilitates comparison with experimental results. In particular, the nanowires are assumed to be cylindrical, and their elastic constants isotropic. The simple analytical model is subsequently validated by performing numerical calculations using realistic nanowire geometries and cubic, anisotropic elastic constants. The comparison confirms that the analytic model is an excellent approximation that greatly facilitates quantitative Raman work, with expected errors in the strain determination that do not exceed 10%. Experimental Raman spectra of a variety of core-shell nanowires are presented, and the strain in the nanowires is assessed using the models described above. It is found that all structures present a significant degree of strain relaxation relative to ideal, fully strained Ge-Si core-shell structures. The analytical models are modified to quantify this strain relaxation.
ContributorsSingh, Rachna (Author) / Menéndez, Jose (Thesis advisor) / Drucker, Jeffery (Committee member) / Ponce, Fernando (Committee member) / Tsen, Kong-Thon (Committee member) / Bennett, Peter (Committee member) / Arizona State University (Publisher)
Created2011
Description
The work described in the thesis involves the synthesis of a molecular triad which is designed to undergo proton coupled electron transfer (PCET) upon irradiation with light. Photoinduced PCET is an important process that many organisms use and the elucidation of its mechanism will allow further understanding of this process and its potential applications. The target compound designed for PCET studies consists of a porphyrin chromophore (also a primary electron donor), covalently linked to a phenol-imidazole (secondary electron donor), and a C60 (primary electron acceptor). The phenol-imidazole moiety of this system is modeled after the TyrZ His-190 residues in the reaction center of Photosystem II (PS II). These residues participate in an intermolecular H-bond between the phenol side chain of TyrZ and the imidazole side chain of His-190. The phenol side chain of TyrZ is the electron transfer mediator between the oxygen evolving complex (OEC) and P680 (primary electron donor) in PSII. During electron transfer from TyrZ to P680*+, the phenolic proton of TyrZ becomes highly acidic (pKa~-2) and the hydrogen is preferentially transferred to the relatively basic imidazole of His-190 through a pre-existing hydrogen bond. This PCET process avoids a charged intermediate, on TyrZ, and results in a neutral phenolic radical (TyrZ*). The current research consists of building a molecular triad, which can mimic the photoinduced PCET process of PSII. The following, documents the synthetic progress in the synthesis of a molecular triad designed to investigate the mechanism of PCET as well as gain further insight on how this process can be applied in artificial photosynthetic devices.
ContributorsPatterson, Dustin (Author) / Moore, Ana L (Thesis advisor) / Gust, Devens (Committee member) / Skibo, Edward B (Committee member) / Arizona State University (Publisher)
Created2011
Description
Introductory programming courses, also known as CS1, have a specific set of expected outcomes related to the learning of the most basic and essential computational concepts in computer science (CS). However, two of the most often heard complaints in such courses are that (1) they are divorced from the reality of application and (2) they make the learning of the basic concepts tedious. The concepts introduced in CS1 courses are highly abstract and not easily comprehensible. In general, the difficulty is intrinsic to the field of computing, often described as "too mathematical or too abstract." This dissertation presents a small-scale mixed method study conducted during the fall 2009 semester of CS1 courses at Arizona State University. This study explored and assessed students' comprehension of three core computational concepts - abstraction, arrays of objects, and inheritance - in both algorithm design and problem solving. Through this investigation students' profiles were categorized based on their scores and based on their mistakes categorized into instances of five computational thinking concepts: abstraction, algorithm, scalability, linguistics, and reasoning. It was shown that even though the notion of computational thinking is not explicit in the curriculum, participants possessed and/or developed this skill through the learning and application of the CS1 core concepts. Furthermore, problem-solving experiences had a direct impact on participants' knowledge skills, explanation skills, and confidence. Implications for teaching CS1 and for future research are also considered.
ContributorsBillionniere, Elodie V (Author) / Collofello, James (Thesis advisor) / Ganesh, Tirupalavanam G. (Thesis advisor) / VanLehn, Kurt (Committee member) / Burleson, Winslow (Committee member) / Arizona State University (Publisher)
Created2011
Description
Source selection is one of the foremost challenges for searching deep-web. For a user query, source selection involves selecting a subset of deep-web sources expected to provide relevant answers to the user query. Existing source selection models employ query-similarity based local measures for assessing source quality. These local measures are necessary but not sufficient as they are agnostic to source trustworthiness and result importance, which, given the autonomous and uncurated nature of deep-web, have become indispensible for searching deep-web. SourceRank provides a global measure for assessing source quality based on source trustworthiness and result importance. SourceRank's effectiveness has been evaluated in single-topic deep-web environments. The goal of the thesis is to extend sourcerank to a multi-topic deep-web environment. Topic-sensitive sourcerank is introduced as an effective way of extending sourcerank to a deep-web environment containing a set of representative topics. In topic-sensitive sourcerank, multiple sourcerank vectors are created, each biased towards a representative topic. At query time, using the topic of query keywords, a query-topic sensitive, composite sourcerank vector is computed as a linear combination of these pre-computed biased sourcerank vectors. Extensive experiments on more than a thousand sources in multiple domains show 18-85% improvements in result quality over Google Product Search and other existing methods.
ContributorsJha, Manishkumar (Author) / Kambhampati, Subbarao (Thesis advisor) / Liu, Huan (Committee member) / Davulcu, Hasan (Committee member) / Arizona State University (Publisher)
Created2011