MATH Graduate Student Conference Papers

Permanent URI for this collection

Browse

Recent Submissions

Now showing 1 - 5 of 30
  • Item
    Clustering and forecasting financial activities at Emprise Bank
    (Wichita State University, 2023-04-14) Rubio Garcia, Fernando; Dao, Mai
    Understanding customers' activities and preferences allows companies to better allocate their resources and streamline services. On this collaborative project, we focus on identifying and extracting useful patterns of patrons' financial transactions and behaviors in the Wichita metropolitan area using a randomly encrypted data sample provided by Emprise Bank. Starting with a preliminary visualization of Wichita-specific geographic, weather, and urbanization details, we further investigate their impacts on customers' preferences to banking services via the Pearson's chi-squared test of independence. Knowledge of such relationships provides us a framework to cluster clients into different groups and observe customer segmentation using K-Means clustering. Finally, we performed time series analysis with the TBATS models to forecast future transactions based on historical information. Our work directly responds to our partner's interests in advancing current and forthcoming branch analytics based on their clientele records.
  • Item
    Measurement of charged pion cross section using newly developed reconstruction tool in the NOvA and dune near detector
    (Wichita State University, 2023-04-14) Roy, Palash Kumer; Yahaya, Abdul-Wasit; Shivakoti, Sushil; Muether, Mathew
    The NOvA and DUNE neutrino experiments are designed to study neutrinos and their interaction properties with matter. Neutrinos, the most abundant massive particle in the universe. They are produced inside the sun, supernovae, galaxies and in nuclear reactors. Neutrinos are electrically neutral particles that come in three types. NOvA and DUNE are both neutrino oscillation experiments which are designed to measure the probability of neutrinos changing types between two detectors that are separated by hundreds of kilometers. Our research group has built a machine learning model to increase the efficiency of determining the point of interaction of particles (Vertex) in the NOvA detector. We plan to use this improved reconstruction to study the likelihood of the production of a charged pion from neutrino interaction. In addition, we are designing a Muon Spectrometer for Phase I of DUNE, which is expected to begin taking data before 2030.
  • Item
    Statistical and classification analysis of Emprise Bank's branches and customers
    (Wichita State University, 2023-04-14) Escamilla, Emilio; Ghazawneh, Sarah; Dao, Mai
    Data-driven insights about customers are important for financial institutions to make informed decisions and to secure economic successes. Such business intelligence is often obtained with a thorough understanding and careful application of statistical analysis techniques. In this collaborative project with Emprise Bank, we employ predictive analytic methods on their randomly encrypted sample data to study financial services offered by branches and utilized by consumers. To understand the connected network of Emprise's branches in the Wichita metropolitan area, we first engage in an exploratory analysis to visualize clients' patterns at branch and ATM sites. Next, we use an analysis of variance to examine the relationships among geographic locations, bank infrastructures, and supported services to observe a balance in customers' businesses in the Northwest and Southeast areas of Wichita, with significantly more credit transactions over all regions. Finally, we employ logistic regression and K-nearest neighbors to predict important services based on patrons' existing physical and electronic footprints at different branches. Future transaction types have the highest prediction accuracies, followed by branch locations where enterprises are to take place.
  • Item
    NLO multijet merging for Higgs production
    (Wichita State University, 2022-04-29) Chen, Tinghua; Figy, Terrance M.
    The discovery of the Higgs boson in 2012 by ATLAS and CMS Collaborations at the Large Hadron Collider (LHC) opens a new era to particle physics, making processes involving Higgs production interesting and essential in many ways. General-purpose Monte Carlo event generators are essential simulation tools for experimental and theoretical physicists. Merging parton shower approximations with next-to-leading order (NLO) matrix element corrections are implemented in the Monte Carlo event generator Herwig 7. We perform simulations of electroweak Higgs boson production in full calculation using the HJets matrix element library. The NLO multijet merging predictions are compared with NLO plus parton shower (NLOPS) matched calculations. This research is under review by The European Physical Journal C. The preprint can be found at arXiv:2109.03730 (2021).
  • Item
    The relationship between preference and performance using three passive exoskeletons during simulated aircraft manufacturing tasks
    (Wichita State University, 2022-04-29) Alqahtani, Haifa; Hakansson, Nils A.; Jorgensen, Michael J.; Jaeger, Adam
    Work-related musculoskeletal disorders (WMSDs) are the top workplace risk factors that impact the health of workers. In Kansas, there were 45,253 reported injuries/illnesses during the 2020 fiscal year. The main contributors to shoulder WMSDs are lifting one's arms and handling heavy tools at or above shoulder level. Passive shoulder exoskeletons are designed to support the arms during overhead work tasks and may aid in reducing the risk of WMSDs. However, exoskeletons are only effective if workers choose to wear them. Therefore, the objectives of this study were to quantify participants' upper-extremity muscle activity reductions as participants performed simulated aircraft manufacturing tasks using three passive exoskeletons and assess correlations between the exoskeleton efficacy and user exoskeleton preferences. To fulfill these objectives, 16 experienced local aircraft manufacturers were recruited to participate in the study. A wireless electromyography (EMG) system was used to record muscle activity levels as the participants performed the simulated work tasks with the exoskeletons. The muscle activity and exoskeleton preference data were analyzed using a multinomial logistic regression model to identify the relationship between participants' muscle activity reduction and exoskeleton preference rankings. The results indicate that the exoskeleton preferences do not correlate with muscle activity reductions. Overall, participants preferred the exoskeleton that provided the least reduction in muscle activity. This study demonstrated that a correlation between the preference and exoskeleton muscle reduction is task dependent, and, therefore, task factors in addition to user preference should be considered in exoskeleton selection.