Smith School of Business Graduate Theses


Recent Submissions

Now showing 1 - 5 of 117
  • Item
    Advanced Analytical Models for Analyzing Financial and Social Data
    Yang, Xingwei; Business; Li, Guang
    Depression is a common mental disorder affecting thoughts, feelings, behaviors, or moods. We develop a methodological language framework that incorporates psychological patterns and states, contextual information, and the influence of social interactions using NLP and ML techniques to detect user-level depression more effectively and accurately from social networks. Social support via social networks can affect mental health outcomes by promoting positive mental states, providing socialization opportunities, and supportive relationships. We define a new group search problem over social networks to find support groups for depressed users. We prove this problem is NP-hard and propose a 2-approximation algorithm. We also conduct experiments using real-world datasets to show that our greedy algorithm can efficiently find groups with shared interests and small proximity. We also use machine learning to enhance business decisions using financial data. We propose a new sampling method specifically suited to the bankruptcy prediction problem. We then calibrate a random forest model as an easy-to-interpret and easy-to-visualize analytic tool to support better investment decision-making by identifying companies at risk of default.
  • Item
    False News: A Digital Conceptualization and Potential Mitigation Using Algorithmic Advice
    Khandoozi (Khan), Seyedali (Ali); Business; Brohman, Kathryn
    Across three conceptual and empirical studies, we respond to urgent calls from all corners of society to address the wicked and universal problem of false messages circulating on the Internet, better known as “fake news”. In the first paper, we problematize the use of the term “fake news”, highlighting issues around its different meanings in prior academic research, its current meaning in the vernacular, and its adequacy to cover the entirety of the phenomena of online falsehoods. Building on existing attempts at addressing the conceptualization problem, we offer our own solution based on the literature on the ontology of digital objects, proposing the concept of “false messages”, of which “false news” is a subset. We also situate this new concept in its broader technical and social context. In the second paper, we shift our attention to mitigating the problem of false news and compare the effects of two algorithmic advisors on individuals’ judgment about news facticity. A large number of algorithms are being developed to identify false news based on the content of news articles (content-based algorithms) or social reaction to news articles (social-based algorithms), which we argue, can act as algorithmic advisors to humans about news facticity. Based on the theory of technology dominance (TTD), Judge-Advisor System (JAS) studies, and computers are social actors (CASA) paradigm, we hypothesize and find some empirical evidence that content-based and social-based algorithmic advisors differ in their ability to influence individuals’ judgments about news facticity. In the final paper, we compare two algorithmic advisors that differ in their source of training data, with one advisor trained using data from a fact-checker with liberal political attitudes and the other trained with data from a fact-checker with conservative political attitudes. Extending the TTD by linking it to similarity-attraction studies, we find different patterns of advice taking from the two algorithmic advisors among US-based Democrats, Republicans, and independents, with Democrats utilizing advice from the algorithmic advisor with liberal training data and Republicans not utilizing advice from either algorithmic advisor, while independents utilized advice from the liberal algorithmic advisor with more nuances compared to the Democrats.
  • Item
    Bai, Lin; Business; Hou, Yu
    In response to growing levels of cyber risk, an increasing number of firms are delegating cyber risk oversight to audit committees. In this dissertation, I examine the role of audit committee (AC) cyber risk oversight in data breach disclosures in SEC filings. By examining breached firms’ disclosure policies, I find that firms with AC cyber risk oversight (1) are more likely to disclose data breaches in SEC filings, and (2) make timelier data breach disclosures in SEC filings. These empirical results support the monitoring role of AC cyber risk oversight. In addition, the effects of AC cyber risk oversight on the likelihood of data breach disclosures in SEC filings are strengthened when the firm is subject to potential regulatory actions. Further to this, AC cyber risk oversight reduces the likelihood that a firm will receive SEC comment letters relating to data breaches, regardless of whether the firm reports a data breach. Finally, all my main results survive in a battery of robustness checks using two-stage regressions and CEM/PSM matching samples, and are further validated in a placebo test. Overall, this dissertation enhances our understanding of how AC cyber risk oversight affects firm’s disclosure policies in relation to data breaches.
  • Item
    Three Essays on Optimal Decisions under Uncertainties or in the Presence of Strategic Customers
    Yousefi, Roozbeh; Business; Wang, Jue
    Three essays on decision-making form this thesis under uncertainties or in the presence of strategic customers. In the first essay, we studied subscription pricing. Subscriptions are agreements between customers and a company that commits to deliver a product or provide a service. We present a continuous-time dynamic pricing model for a monopolist offering a fixed-term subscription contract, without per-use charges and access limits, to strategic customers whose utility is affected by the number of subscribers. We formulate the monopolist’s problem in terms of optimal control, derive its optimality conditions, and study the structure of the stationary optimal solution. We study the transient and steady states of the problem and show that the firm can capitalize on the strategic behavior of customers. We also observe that the firm is better off not offering subscriptions for myopic customers or customers with a low evaluation of the service. Finally, we demonstrate the robustness of the optimal pricing results in a setting in which we relax significant assumptions. The second essay considers the joint monitoring and learning of a partially observable dynamic system in which all parameters are unknown. This problem is motivated by the monitoring of cryptic threatened species. The objective is to detect a change in the hidden state while simultaneously learning about the system dynamics and observability. We formulate the problem in the framework of Bayes-adaptive partially observable Markov decision processes (Bayes-adaptive POMDP). A distinguishing feature is that decisions are made under the coupling of state uncertainty, parameter uncertainty, and state dynamics. We identify a low-dimensional sufficient statistic and reformulate the dynamic program in three dimensions. We fully characterize the optimal policy structure. The advantages of the monitoring-while-learning strategy are demonstrated with a case study on the Sumatran tiger conservation in Indonesia. The third essay investigates a dynamic pricing problem in a stochastic setting and proves that the scaled version of the problem has a fluid limit. The deterministic problem has continuous variables and equations. The optimal control tool is a perfect method to analyze similar continuous-time problems, and it delivers results that are not available in a stochastic setting efficiently.
  • Item
    Harnessing Digital Technologies for Entrepreneurial Innovation
    Li, Ting; Business
    This thesis, organized in a 3-paper format, explores the potential of digital technologies to support entrepreneurial innovation using a capability-based perspective. The fast emergence of powerful, readily accessible, and affordable new technologies has transformed the processes and outcomes of entrepreneurial innovation. Both startups and business incubators that support startups, as a result, require an up-to-date and nuanced understanding of how digital technologies can be leveraged to create business value in their respective contexts. In the first paper, a multilevel framework is inductively developed to explain how digitally-enabled influential factors at the ecosystem level shape innovation environments through business incubators; how startups carry out digitally-enabled entrepreneurial actions; and how business incubators and startups jointly form digitally-enabled intra-incubator action patterns to co-create innovation outcomes at the ecosystem level. Findings of this study confirm the saliency of digital enablement at all three analytical levels (i.e., entrepreneurial ecosystems, business incubators, and startups) and throughout the innovation processes. The second paper takes a mixed-method approach to explore the value of digital capabilities to startups. Using case data collected from 65 Canadian startups, this study inductively identifies four types of digital capabilities that are commonly developed and used by early-phase startups: digital platform capability, digital infrastructure capability, digital adaptation capability, and digital knowledge capability. Drawing on the insights derived, fuzzy-set qualitative comparative analysis was conducted to examine high-performing digital capability configurations for startups at the business validation and transition stages. The results show that the high-performing configurations differ from validation to transition, and that digital startups and their less-digital counterparts rely on different digital capability configurations to succeed. The third paper investigates further a key finding from the first paper – that is, today’s business incubators are being transformed from siloed support infrastructure to active resource orchestrators in entrepreneurial ecosystems. Specifically, this study examines the influence of dynamic digital capabilities on both the agility and performance of publicly-funded business incubators. The research framework is tested and supported using survey data. It is found that resource orchestration capabilities fully mediate the impact of dynamic digital capabilities on incubator performance, and partially mediate the latter’s impact on incubator agility.