<p> Big data queries are increasing in complexity and the performance of data analytics is of growing importance. To this end, Big Data on high-performance computing (HPC) infrastructure is becoming a pathway to high-performance data analytics. The state of performance studies on this convergence between Big Data and HPC, however, is limited and ad hoc. A systematic performance study is thus timely and forms the core of this research. </p><p> This thesis investigates the challenges involved in developing Big Data applications with significant computations and strict latency guarantees on multicore HPC clusters. Three key areas it considers are thread models, affinity, and communication mechanisms. Thread models discuss the challenges of exploiting intra-node parallelism on modern multicore chips, while affinity looks at data locality and Non-Uniform Memory Access (NUMA) effects. Communication mechanisms investigate the difficulties of Big Data communications. For example, parallel machine learning depends on collective communications, unlike classic scientific simulations, which mostly use neighbor communications. Minimizing this cost while scaling out to higher parallelisms requires non-trivial optimizations, especially when using high-level languages such as Java or Scala. The investigation also includes a discussion on performance implications of different programming models such as dataflow and message passing used in Big Data analytics. The optimizations identified in this research are incorporated in developing the Scalable Parallel Interoperable Data Analytics Library (SPIDAL) in Java, which includes a collection of multidimensional scaling and clustering algorithms optimized to run on HPC clusters. </p><p> Besides presenting performance optimizations, this thesis explores a novel scheme for characterizing Big Data benchmarks. Fundamentally, a benchmark evaluates a certain performance-related aspect of a given system. For example, HPC benchmarks such as LINPACK and NAS Parallel Benchmark (NPB) evaluate the floating-point operations (flops) per second through a computational workload. The challenge with Big Data workloads is the diversity of their applications, which makes it impossible to classify them along a single dimension. Convergence Diamonds (CDs) is a multifaceted scheme that identifies four dimensions of Big Data workloads. These dimensions are problem architecture, execution, data source and style, and processing view. </p><p> The performance optimizations together with the richness of CDs provide a systematic guide to developing high-performance Big Data benchmarks, specifically targeting data analytics on large, multicore HPC clusters.</p>

<p> Researchers, epidemiologists and policy analysts rely on birth certificate data for population health surveillance on maternal and infant health trends. Inaccurate data collection can misguide and distort important conclusions for those that depend on these data. Registered obstetric and neonatal nurses are in the unique position to positively impact quality data collection for birth reporting. </p><p> Dr. Benner&rsquo;s Novice to Expert Nursing Theory (1982) and Dr. Knowles&rsquo; Adult Learning Theory (1980) provided a conceptual framework for the development of a quasi-experimental pilot to: 1. Investigate the obstacles and barriers that impact the obstetric and neonatal registered nurses&rsquo; ability to collect and document information for birth reporting; 2. Assess the influence of a new eLearning training system on obstetric and neonatal registered nurses&rsquo; knowledge and level of satisfaction. </p><p> Twenty-three registered nurses from the Association of Women's Health, Obstetric and Neonatal Nurses (AWHONN) with experience in labor and delivery, maternal and child health completed an eLearning system and a classic pre-test/post-test questionnaire that included Likert-scale assessment questions. Inadequate training in best practices was the main obstacle to birth reporting. Paired-sample t-test of pre/posttest scores were statistically significant. Findings suggested eLearning may serve as an effective educational tool for birth reporting. Further investigation in an expanded study is warranted including consideration for competency-based eLearning, and engaging healthcare organizations and nurse educators. </p>

<p> Roadway systems in the United States have become huge assets that need massive resources to maintain and operate. To meet the long-term performance goal, government agencies developed pavement management systems (PMSs) to help them manage roadway assets effectively with limited resources. Currently, some PMSs in the United States have been designed for two types of pavements: asphalt and concrete. The composite pavement, another pavement type, which is the result of concrete pavement rehabilitations and constructed with an asphalt surface layer over a concrete base, was treated as asphalt. However, the literature review indicates that compared to asphalt pavements, composite pavements perform differently and have different dominant distresses. In addition, as the amount of composite pavements increases, it is necessary to investigate them independently to incorporate more accurate information into the PMS. Therefore, the goal of this research is to improve and to expand the PMS with an additional pavement type: composite pavements. To achieve this goal, the PMS managed by the North Carolina Department of Transportation (NCDOT) was used as a case study, and several objectives were accomplished in this research: 1) to identify composite pavements and generate the raw data based on the construction history; 2) to clean the raw data and mitigate errors using statistical methods and engineers&rsquo; experiences; 3) to develop nonlinear models to describe dominant distresses and pavement performances; 4) to propose quantile regression (QR) models to predict pavement performances; and 5) to investigate the pavement treatment effectiveness by exploring performance index jumps. </p><p> Based on findings of this research, it was concluded that the automated data were more consistent with engineers&rsquo; experience and revealed more information than the windshield data; longitudinal cracking and transverse cracking were found to be the dominant distresses in composite pavements, followed by alligator cracking and raveling; Interstate composite pavements deteriorated faster than both US and NC composite pavements, and NC composite pavements had the slowest deterioration rate; QR models can be used as a new prediction method of pavement performances at both the project and the network levels; in general the &ldquo;Resurfacing&rdquo; treatment was more effective than the &ldquo;Chip Seal&rdquo; treatment; and The average service life of asphalt and composite pavements were similar, but composite pavements have a smaller variation of service lives than that of asphalt pavements.</p><p> It was recommended that the automated data should be used in future PMS related research projects, due to its better data quality, and because of the robust performance of QR models at both network and project levels, QR models should be incorporated in the future PMS.</p><p> In summary, this research expanded the existing NCDOT PMS with composite pavements, proposed systematic methods to improve the quality of performance data, enriched the diversity of prediction models by exploring potentials of QR models, and investigated the effectiveness of pavement treatments. Essentially, transportation agencies can use the findings of this research to make informative investment decisions.</p>

<p> The author analyzed survey responses from fifteen adjunct faculty members teaching in applied graduate programs to examine the language they used to discuss information literacy in conjunction with their own research and the teaching of graduate college students. The adjuncts are industry professionals, experts in their respective fields. The college, Goucher College, is a small private liberal arts college established in 1885, in the state of Maryland. Three research questions guided the work - how do faculty come to understand information literacy, what language do they use to describe it, and how do they transfer this knowledge to their students.</p>

No description available.

Survey responses from 1,962 students attending Florida State University during the Spring 2007 semester were analyzed for a preferred sensory learning mode, number and variety of study methods, a match between sensory learning mode and sensory preference within the study habits, and passive versus deep learning processes. The students were found to demonstrate a sensory learning mode preference, but did not report using study habits that matched this sensory mode. Most of the students were found to employ too few study methods to allow for the development of true study strategies, were found to use memorization as a primary study method, and to utilize passive learning rather than deep learning processes. / A Thesis submitted to the College of Information in partial fulfillment of the requirements for the degree of Master of Science. / Degree Awarded: Fall Semester, 2007. / Date of Defense: October 12, 2007. / College Students, Passive Learning, Study Strategies, Sensory Learning, Learning Mode / Includes bibliographical references. / Gary D. Burnett, Professor Directing Thesis; Norbert M. Seel, Committee Member; Paul F. Marty, Committee Member.

The accelerated development of information technology, especially the rapid growth of the Web, is changing the circumstances and consequently the structures and processes of scholarly communication. Since the Internet as a powerful communication medium has radically improved the efficiency of communication, scholarly communicative activities are increasingly being conducted over the Internet: interacting with peers, searching for information, publishing research results, etc. As a result, there is renewed interest in the study of scholarly communication to see the types of communication that are taking place and the similarities to what we have come to expect from print based communication. New data sources and tools for scholarly communication research are becoming increasing available on the Web as well. These data and tools have opened up the possibility of new topics of inquiry applying new methods leading to new theories (Borgman, 2002; Zhao & Logan, 2002). The present study explores this possibility through an author citation analysis of scholarly communication patterns in the extensible Markup Language (XML) research field using data both from the Web as indexed by Research Index (http://www.researchindex.com) and from print journals as indexed by the Institute for Scientific Information (ISI)s Science Citation Index (SCI). A series of citation analyses including author visibility analysis and author co-citation analysis have been conducted, and comparisons of results from the two data sources have been carried out controlling for data scope and citation counting method respectively, to identify the similarities and differences between Web-based and print journal-based scholarly communication as revealed by citation analysis and to reveal the capacity of scientific papers published on the Web along with existing citation indexing tools for Web publications as an alternative data source to the ISI databases for citation analysis studies. Meanwhile, publications and characteristics of three groups of highly visible authors have been examined and compared authors highly visible both on the Web and in journals, those only in journals, and those only on the Web, to ascertain the possible contributing factors to the differences identified from the two data sources. Based on these, the present study also explores possible improvements of data sources and tools on the Web and the requirements for a problem solving environment (PSE) for scholarly communication research. XML is one of the most promising research areas and also one of the research fields that has taken an early move to online publishing. Through a series of citation analyses of the XML research field, the present study identifies the ideas and thinkers that have influenced the conceptual development within the field, the sub-areas in this field that appear to be more active, and the interrelationship of the sub-areas and of core researchers. Findings from this study indicate that the XML research field is a coherent research field in which scholars are well interconnected through intellectual and social ties. Researchers in this field are communicating research results heavily on the Web, especially results from studies at the research front. The two groups of XML scholars who actively publish on the Web or in journals share very few publications, and are concerned with different issues. While all study XML related standards or specifications and XML database design and implementation, research on XML applications is a focus only in journals, and research into the Semantic Web and programming for and processing of XML data is better represented on the Web. It appears that while emerging specialties such as the Semantic Web are more visible on the Web, new trends in long-existing specialties such as Software agents are well represented in journals. However, these two groups have similar citing behavior and their collective view of author visibility and the intellectual structure of the research field is very similar as well. This is indicated by the very high correlation between author ranking by number of citations resulting from ResearchIndex and that from SCI when the same citation counting method was used. This is also suggested by the similar relative locations of specialties, to which both groups have contributed, on the maps resulting from Multi-Dimensional Scaling (MDS) of the two data sources. Conclusions can be drawn from the findings that evaluation of scholars and examination of intellectual structures based on the collective view of citers on the Web should be considered as equally valid as those based on citers view in journals, provided the discipline being studied is well-published on the Web. In order to gain a complete picture of the scholarly communication patterns in the XML research field, multiple data sources should be used rather than only the ISI databases or ResearchIndex. These findings also seem to evidence a two-tier system in scholarly communication that is believed by some scholars to be a probable future model of the scholarly communication system (Poultney, 1996; van Raan, 2001). In this model, the first tier is a free space which represents the scholarly enterprise in real time and is most likely to feature free Web-based publications, while the second tier is the world of more formal publications that is most likely to continue to be dominated by journals (van Raan, 2001, p. 61). In other words, in this model, research would largely be initially reported on the Web to obtain priority and fast recognition and then gradually distributed through other more formal channels such as journals to gain formal acceptance. Data sources and tools increasingly available on the Web have opened up the possibility of new topics of inquiry applying new methods leading to new theories. However, they currently do not cover as many disciplines and are not as easy to use as the ISI databases. These are some of the aspects in scholarly communication systems that need to be improved and to which citation analysis can contribute. A well designed problem solving environment (PSE) for scholarly communication research can be a solution. The preliminary exploration of such a PSE in the present study is a start and hopefully will lead to a full discussion in the future. / A Dissertation submitted to the School of Information Studies in partial fulfillment of the requirements for the degree of Doctor of Philosophy. / Degree Awarded: Summer Semester, 2003. / Date of Defense: April 21, 2003. / Efficiency Of Communication / Includes bibliographical references. / Gary Burnett, Professor Directing Thesis; Lawrence C. Dennis, Outside Committee Member; Elisabeth Logan, Committee Member; Kathleen Burnett, Committee Member; Benjamin Keith Belton, Committee Member.

This study examined the effectiveness of the young adult library specialist and the generalist librarian in reference to user satisfaction scores given by young adult clientele and based on their evaluation of their local public library. The performance measure of user satisfaction served to infer the effectiveness of public library service to this clientele. A questionnaire assessed attitudes of young adult library users in two large urban public libraries, one of which employs a young adult specialist and one of which serves this clientele with generalists. The questionnaire collected minimal background data on the respondents, measured satisfaction scores and tested the null hypothesis by way of the Critical Value of Variance Ratio F within the t-Test for Two Independent Samples. / Based on the findings of the study, the following conclusion was reached: a specialist young adult librarian makes a significantly positive difference in user satisfaction scores for the following factors: a feeling of welcome to the young adults who frequent their local public library, use of interlibrary loan, knowledge of hotline information, use of programming and activities, interest in a popular materials collection and enjoyment of related booklists on young adult reading, and knowledge of a area for playing games. A library which employs a young adult librarian is more effective than a library which does not for each of these services. Other findings of this study were that generalist librarians make a significantly positive difference in user satisfaction scores for the following factors: knowledge and interest on the topics of sexuality, sexual and physical abuse, and pregnancy. / The major recommendation of this study is that young adult public library users be provided with a young adult specialist and a room in the public library dedicated to their special needs. These needs include improved collection strength and availability to information related to sexuality, sexual or physical abuse, and pregnancy. / Other recommendations and suggestions for further study are also included. / Source: Dissertation Abstracts International, Volume: 53-10, Section: A, page: 3400. / Major Professor: F. William Summers. / Thesis (Ph.D.)--The Florida State University, 1992.

Electronic bulletin boards have become more popular with the advent of low cost computers, easily available communication channels, and large network systems. With its public posting of messages, relatively large audiences can read and provide immediate feedback. Network organization provides a hierarchical system to give users access to thousands with like interests across the country, or even the world. / A random sample of bulletin board users was selected from the two major national bulletin board services, CompuServe$\sp\circler$ and Prodigy$\sp\circler,$ and interviewed asynchronously, on-line, using electronic mail. / This study characterized this bulletin board audience focusing on the issues of (1) adoption--what characteristics bulletin board users possess, and (2) social impacts--how bulletin board adoption affects the use of other communication media. / Results of the on-line survey supported adoption hypotheses suggesting bulletin board users are higher educated, more wealthy, and work in more prestigious occupations the average American. Rejected was a similar lower age hypothesis. Open-ended responses to the question "why use boards?" were content analyzed, with selections made to categories in five areas showing a typology of use for this medium. Hypotheses suggesting a reduction of consumption of the media television, books, telephone use, and personal letter writing were supported. Rejected were suggestions of lower magazine reading, face-to-face conversations, and other on-line communication. / The findings also showed significant differences in the CompuServe and Prodigy systems along several of the above variables, in addition to quantity of use, message rate, and family usage patterns. / Source: Dissertation Abstracts International, Volume: 53-09, Section: A, page: 3029. / Major Professor: C. Edward Wotring. / Thesis (Ph.D.)--The Florida State University, 1992.

Although there are numerous studies, conferences, and journal articles devoted to the topic of library education, extensive debates remain regarding both the purpose and effectiveness of library education as a tool in helping prepare entry-level professional librarians. A major part of the debate is the question of whether, in general, library schools are meeting the demands of the job market for librarians and information specialists. The purpose of this study was to investigate the preparation of entry-level research librarians. It will describe the relationship between formal library education and work in research libraries from the perceptions of entry-level librarians and their supervisors. The main goal of the study is to help library and information studies schools to prepare future research librarians and equip them with in-demand knowledge and skills. The results of the study would be of benefit for administrators and educators at library and information studies schools who are concerned with preparing MLIS students to work in research libraries. In order to answer the five research questions designed for this study, two web based surveys were used to collect data from 45 entry-level research librarians in Florida, and 18 heads of departments at the same libraries. Descriptive and inferential statistics were used for data analysis. Findings of the study validated 28 required competencies for entry-level research librarians. Perceptions of research librarians indicated that schools of library and information studies in the U.S. have helped their recent graduates, who work in research libraries, in acquiring 64% of the required competencies. In addition, findings indicated that increasing internship requirements and adding more management instruction are the most important methods for library schools to consider in preparing future research librarians. Finally, the study elicited the perceptions of both groups regarding the expected required competencies for entry-level research librarians from 2005 to 2010. Recommendations for library schools regarding the educational preparation of entry-level research librarians are provided. / A Dissertation submitted to the School of Information Studies in partial fulfillment of the requirements for the degree of Doctor of Philosophy. / Degree Awarded: Spring, 2003. / Date of Defense: April 14, 2003. / Research Libraries / Includes bibliographical references. / Jane Robbins, Professor Directing Dissertation; Peter Garretson, Outside Committee Member; F. William Summers, Committee Member; Thomas Hart, Committee Member.