INFORMS 2021 Program Book

INFORMS Anaheim 2021

TD14

3 - Data Tracking under Competition Ilan Morgenstern, Stanford University, Stanford, CA, United States, Kostas Bimpikis, Daniela Saban

2 - Data Mining and Business Analysis for Efficient Search Engine Marketing Arif Ansari, University of Southern California, Los Angeles, CA, 90089-0809, United States In this presentation, we will provide a novel approach to find Search Engine Marketing keywords using multiple datamining methods, like clustering, text explorer, Tree Maps, Probabilistic Simulation, Estimation of Customer Acquisition Cost etc., This presentation will show the confluence of Business Analysis, Analytics and Data Mining to get Competitive Edge. 3 - Introducing the Multivariate Returns to Scale Technology Dariush Khezrimotlagh, Pennsylvania State University Harrisburg, Middletown, PA, 17057-4846, United States In this study, the motivation and construction of the multivariate returns to scale technology (MRTS) are presented. A linear programming data envelopment analysis (DEA) model is suggested as a data-driven tool to measure the corresponding production function and measure the corresponding efficiency score of units. The corresponding technology includes the corresponding variable returns to scale (VRS) technology and is a subset of the constant returns to scale (CRS) technology. As a result, the proposed score is neither less than that of the corresponding CRS DEA model nor greater than that of the corresponding VRS DEA model. The method should be considered for real-life applications when datasets include more than one input and/or more than one output. The process to justify the returns to scale for a dataset is also illustrated. 4 - Irrational Exuberance: Correcting Bias in Probability Estimates Bradley Rava, University of Southern California, Los Angeles, CA, United States, Peter Radchenko We consider the common setting where one observes probability estimates for a large number of events. Selecting events corresponding to the most extreme probabilities can result in systematically underestimating the true level of uncertainty. We develop an empirical Bayes approach “Excess Certainty Adjusted Probabilities” (ECAP), using a variant of Tweedie’s formula, which updates probability estimates to correct for selection bias. ECAP directly estimates the score function associated with the probability estimates, so it does not need to make any restrictive assumptions about the prior on the true probabilities. ECAP also works well in settings where the probability estimates are biased. We demonstrate through theoretical results and empirical analysis that ECAP can provide significant improvements over the original probability estimates. 5 - Integrating INFORMS CAP/aCAP Into Your Academic Program Matthew A. Lanham, Purdue University, Lafayette, IN, 47905- 4803, United States We discuss how we are making a proactive attempt to get every future graduate of Purdue’s M.S. in Business Analytics & Information Management (BAIM) program and many program alumni to become INFORMS Certified Analytics Professionals (CAPs/aCAPs). We discuss why most programs have not achieved this, why we believe we can, how we are going about it, and challenge any academic program to partner with us to create additional incentives to push our graduates and program outcomes to the next level. TD14 CC Room 201B In Person: Data and Modeling Applications in Response to COVID-19 Pandemic General Session Chair: Yuan Zhou, University of Texas at Arlington, Arlington, TX, 76019-1000, United States 1 - An Agent-based Modeling Approach for Planning COVID-19 Reopening Activities Yuan Zhou, University of Texas at Arlington, Arlington, TX, 76019-1000, United States With mass distributions of vaccines, COVID-19 cases and fatalities have been reduced significantly in the US. However, it remains unclear when communities can return to the pre-pandemic normalcy completely, where the effectiveness of reopening strategies plays a critical role at both community- and individual-level. To derive an effective reopening plan, this study develops an agent-based simulation. The proposed model mimics the underlying transmission dynamics under different intervention scenarios and enables joint investigations of the timing and magnitude of lifting interventions used currently, such as social distancing precaution, mandatory mask wearing order, and online working requirement.

We explore the implications of data tracking technologies that enable firms to collect consumer data and use it for price discrimination. We find that the absence of data tracking may lead to a decrease in consumer surplus, even if consumers are myopic. Importantly, this result relies on competition: consumer surplus is higher with data tracking only when multiple firms offer substitutable products to consumers. Our results contribute to the debate of whether to regulate firms’ use of data tracking technologies by illustrating that their effect on consumers depends not only on their level of sophistication, but also on the degree of competition in the market.

TD12 CC Room 304D In Person: New Frontiers in Behavioral OM General Session

Chair: Evgeny Kagan, Johns Hopkins Carey Business School, Johns Hopkins Carey Business School, Baltimore, MD, 21202-4673, United States 1 - Seeing the Bigger Picture? Ramping Up Production with the Use of Augmented Reality Enno Siemsen, University of Wisconsin-Madison, Wisconsin School of Bus., Madison, WI, 53706-1324, United States, David Wuttke, Ankit Upadhyay, Alexandra Wuttke-Linnemann Firms increasingly use augmented reality (AR) devices to improve their production ramp-up processes. Our field experiment provides empirical evidence related to the strengths and weaknesses of AR in the ramp-up of production. When faced with a new task, workers instructed by AR smart glasses achieve a 74% higher flow rate compared with a control group. However, workers that use AR glasses consistently perform 21% slower than the control group when both groups repeat the task, without either AR or paper-based instructions. After the devices have been removed, workers instructed based on paper improve their productivity faster through learning than those instructed by AR. In addition, the former group suggests better process improvements than the latter. 2 - Understanding Donors’ Preferences in Charitable Giving Hasti Rahemi, University of Colorado Boulder, Boulder, CO, United States, Sebastian Villa, Gloria Urrea Charities depend on donors for funding to run their operations. However, it is not clear how donors’ preferences affect donors’ contributions. We propose that to fully understand donors’ decisions, it is necessary to account for two types of donors’ preferences: their predilection for programs and their own self-serving bias. We investigate these two types of preferences in an online experiment with over 350 participants. Our results provide guidance for charities to improve their fundraising strategies. TD13 CC Room 201A In Person: Confluence of Data Mining and Business Analysis General Session Chair: Arif Ansari, University of Southern California, Los Angeles, CA, 90089-0809, United States Co-Chair: Dariush Khezrimotlagh, Pennsylvania State University Harrisburg, Middletown, PA, 17057-4846, United States 1 - Visual Programming: Teaching for Business Analytics Best Practices Dursun Delen, Professor & Research Director, Oklahoma State University, Tulsa, OK, United States A proven way to make learning of the foundational concepts and best practices more intuitive for students is to utilize a visual modeling and workflow-driven analytics platform. The goal is to reduce the syntactic nature of data science so that more time and mental capacity can be spent on concepts. KNIME Analytics Platform (a free and open-source software environment) is an excellent candidate for such an intuitive teaching, learning, and practicing tool. The current presentation will provide evidence and a short tutorial to illustrate the ease use of the KNIME Analytics Platform.

129