Theses and Dissertations - Department of Information Systems, Statistics & Management Science

Permanent URI for this collection

https://ir.ua.edu/handle/123456789/114

Browse

Browsing Theses and Dissertations - Department of Information Systems, Statistics & Management Science by Subject "Computer science"

Now showing 1 - 4 of 4
  • Thumbnail Image
    Item
    Advancing evidence-based practice in systems development: providing juried knowledge to software professionals
    (University of Alabama Libraries, 2014) Hassler, Edgar E.; Hale, David P.; University of Alabama Tuscaloosa
    The concept of utilizing information derived from carefully crafted scientific research to optimize the efficiency and/or effectiveness of a practice is by no means a new idea. For over two thousand years physicians, scientists, and business professionals have relied on evidence to improve decision making. The advances made over last 50 years with regard to Information Systems (IS), and the proliferation of technology, set the stage for a new paradigm in the use of information in practice commonly referred to as Evidence-Based Practice (EBP). Originating in the field of medicine, the EBP paradigm has been adopted in many of the healthcare domains and spread to other domains such as education, management, and computer science. In the spirit of Fredrick Taylor's The Principles of Scientific Management (1914), the collection of essays presented in this work endeavor to advance the use of empirically based, juried evidence for decision making in a business context. The specific context selected is the domain of SE - a key component in a technology laden world. Within the SE domain, the essays address three objectives in the advancement of the EBSE paradigm by: 1. Mapping the research completed to date regarding the implementation of EBP in the SE domain to identify gaps and opportunities in the research. 2. Identifying the barriers deemed most important by the members of the SE research community who conduct systematic literature reviews in support of EBSE. 3. Developing the use of algorithmic techniques as a discriminant function in the selection process of the systematic review methodologies. Together, the collection of essays represent a line of inquiring within a broader research stream concerning the implementation of EBP - a modern version of Taylors work - within the SE domain. The collection of essays provides valuable insights concerning the status of EBSE and its literature, the problems associated with secondary research under the paradigm, and the basis for a discrimination function designed to assist in resolving a key issue for those seeking guidance in academic literature.
  • Thumbnail Image
    Item
    GA-Boost: a genetic algorithm for robust boosting
    (University of Alabama Libraries, 2012) Oh, Dong-Yop; Gray, J. Brian; University of Alabama Tuscaloosa
    Many simple and complex methods have been developed to solve the classification problem. Boosting is one of the best known techniques for improving the prediction accuracy of classification methods, but boosting is sometimes prone to overfit and the final model is difficult to interpret. Some boosting methods, including Adaboost, are very sensitive to outliers. Many researchers have contributed to resolving boosting problems, but those problems are still remaining as hot issues. We introduce a new boosting algorithm "GA-Boost" which directly optimizes weak learners and their associated weights using a genetic algorithm, and three extended versions of GA-Boost. The genetic algorithm utilizes a new penalized fitness function that consists of three parameters (a, b, and p) which limit the number of weak classifiers (by b) and control the effects of outliers (by a) to maximize an appropriately chosen p-th percentile of margins. We evaluate GA-Boost performance with an experimental design and compare it to AdaBoost using several artificial and real-world data sets from the UC-Irvine Machine Learning Repository. In experiments, GA-Boost was more resistant to outliers and resulted in simpler predictive models than AdaBoost. GA-Boost can be applied to data sets with three different weak classifier options. We introduce three extended versions of GA-Boost, which performed very well on two simulation data sets and three real world data sets.
  • Thumbnail Image
    Item
    On the use of transformations for modeling multidimensional heterogeneous data
    (University of Alabama Libraries, 2019) Sarkar, Shuchismita; Melnykov, Volodymyr; University of Alabama Tuscaloosa
    The objective of cluster analysis is to find distinct groups of similar observations. There are many algorithms in literature that can perform this task and among them model based clustering is one of the most flexible tools. Assumption of Gaussian density for mixture components is quite popular in this field of study due to it’s convenient form. However, this assumption is not always valid. This thesis explores the use of various transformations for finding clusters in heterogeneous data. In this process, the thesis also attends to several data structures such as vector-, matrix-, tensor-, and network-valued data. In the first chapter, linear and non-linear transformations are used to model heterogeneous vector-valued observations when the data suffer from measurement inconsistency. The second chapter discusses an extensive set of parsimonious models for matrix-valued data. In the third chapter a methodology for clustering skewed tensor-valued data is developed and it is applied for analyzing remuneration of professors in American universities. The fourth chapter focuses on network-valued data and a novel finite mixture model addressing the dependent structure of network data is proposed. Finally, the fifth chapter describes the functionality of a R package “netClust” developed by the author for clustering unilayer and multilayer networks following the methodology proposed in Chapter four.
  • Thumbnail Image
    Item
    Specification theory, patterns, and models in information systems domains: an exploratory investigation
    (University of Alabama Libraries, 2010) Woolridge, Richard William; Hale, David P.; University of Alabama Tuscaloosa
    Application and project domain specifications are an important aspect of Information Systems (IS) development. Observations of over thirty IS projects suggest dimly perceived structural patterns in specifications that are unaccounted for in research and practice. This investigation utilizes a theory building with case studies methodology to elucidate some of these patterns. As prerequisites to pattern identification, this investigation identified a theoretically and empirically grounded static model of specification that establishes specification context and structure and a theoretically and empirically grounded dynamic model of specification that establishes the principles of specification emergence and evolution. Using these models as a foundation, this investigation synthesized a specification pattern model from four research disciplines and confirmed the specification pattern model for physical objects in case data. An additional specification pattern model for physical actions was also identified in the case data. The confirmation process found that the physical object and action models could not be extended to abstract informational objects or actions. The findings of this investigation answer a call for research of the application domain, advance understanding of IS requirements inadequacy and volatility, advance ontological research to include a mechanism to integrate static and dynamic dimensions, and provides avenues of study to improve understanding of the structure and dynamics governing stakeholder collaboration. In addition, this investigation suggests criteria to judge, as well as theories, models, and patterns for, a class of IS that is persistently adaptive.