Research

Research Summary

The bulk of my research combines machine learning and optimization techniques to build decision support systems that improve companies’ business processes and information flow.

Most of my work aims at maximizing efficiency and social fairness in medical appointment scheduling. My research finds that employing machine learning to predict patient no-shows leads to more efficient clinic operations. However, it also finds that doing so may cause patients belonging to more vulnerable racial groups to experience longer wait times at the clinic. In collaboration with an interdisciplinary team of experts in operations management, information systems, ethics, and public policy, I overcome this problem by developing appointment scheduling methodologies that minimize racial disparity while maximizing clinic efficiency.

Machine learning and optimization are common themes in my other research as well, which focuses on developing general-purpose methodologies applicable to a variety of domains. For example, I developed visualization and text mining methodologies to analyze the evolution of peer-produced artifacts (e.g., Wikipedia articles), relational data mining techniques to support the drug discovery process at pharmaceutical companies, and metaheuristic techniques to tackle hard optimization problems.

As I strive to produce research of high practical value and positive social impact, I have been collaborating with several organizations, such as a healthcare organization dedicated to Black women, a nonprofit mental health center, a community legal center, and a pharmaceutical company.

Highlights

  • 25 refereed publications (14 journals, 10 conference proceedings, 1 book chapter) including top venues in Information Systems, Operations Management, and Operations Research (e.g., MIS Quarterly, INFORMS Journal on Computing, Manufacturing & Service Operations Management, Production and Operations Management).

  • My research on racial disparities in health care cited in a United Nations report, presented at the UN Human Rights Council

  • Selected by AACSB in the “2021 Innovations That Inspire” (one of only 24 projects selected among 890 AACSB accredited business schools)

  • Research featured or mentioned by multiple media outlets (such as Forbes and WIRED)

  • One of only five academics to be invited in 2019 by the National Academies of Sciences, Medicine, and Engineering to advise the US Department of Veteran Affairs on their appointment scheduling system (see here)

  • Two of my papers used in doctoral seminars at other institutions

Journal Articles

  1. Harris, S., K. Hicklin, M. Samorani. M.A. Santoro. Reducing Racial Disparities in Medical Appointment Scheduling by Eliminating Priority Appointment Slots (working paper)

  2. Shanklin, R., M. Samorani, S. Harris, M.A. Santoro. Ethical Redress of Racial Inequities in AI: Lessons from decoupling Machine Learning from Optimization in medical appointment scheduling, Philosophy and Technology, 35(96) (paper).

  3. Samorani, M., R. Bala, R. Jacob, S. He. 2022. A Software Package and Data Set for the Personal Protective Equipment Matching Problem During Covid-19. INFORMS Journal on Computing (featured article) (paper, GitHub page)

  4. Samorani, M., S. Harris, L.G. Blount, H. Lu, M.A. Santoro. 2021. Overbooked and Overlooked: Machine Learning and Racial Bias in Medical Appointment Scheduling. Manufacturing & Service Operations Management (paper, video)

  5. Harris, S., M. Samorani. 2020. On Selecting a Probabilistic Classifier for Appointment No-show Prediction (the authors contributed equally to this article). Decision Support Systems, p. 113472.

  6. Lu, H., X. Chen, Q. Liu, M. Samorani, G. Song, X. Chen, Y. Yang. 2020. Stochastic Workflow Authorizations with Queueing Constraints. IEEE Transactions on Dependable and Secure Computing. doi: 10.1109/TDSC.2020.3026296.

  7. Samorani, M., L.G. Blount. 2020. Machine Learning and Medical Appointment Scheduling: Creating and Perpetuating Inequalities in Access to Health Care. The American Journal of Public Health, 110(4), pp. 440–441 .

  8. Arazy, O., A. Lindberg, M. Rezaei, M. Samorani (the authors contributed equally to this article). 2021. The Evolutionary Trajectories of Peer-Produced Artifacts: Group Composition, the Trajectories' Exploration, and the Quality of Artifacts. MIS Quarterly, 44(4).

  9. Samorani, M., A. Alptekinoglu, P. Messinger. 2019. Product Return Episodes in Retailing. Service Science, 11(4), pp.263-278.

  10. Soltani, M., Samorani, M. and Kolfal, B., 2019. Appointment scheduling with multiple providers and stochastic service times (online supplement). European Journal of Operational Research, 277(2), pp.667-683.

  11. Samorani, M., Y. Wang, Y. Wang, Z. Lv, F. Glover. 2019. Clustering-driven evolutionary algorithms: an application of path relinking to the quadratic unconstrained binary optimization problem. Journal of Heuristics, 25(4-5), pp.629-642.

  12. Rezaei, M., Cribben, I. and Samorani, M., 2018. A clustering-based feature selection method for automatically generated relational attributes. Annals of Operations Research, pp.1-31.

  13. Samorani, M., S. Ganguly. 2016. Optimal Sequencing of Unpunctual Patients in High-Service Level Clinics. Production and Operations Management, 25(2) 330-346.

  14. Samorani, M., L. LaGanga. 2015. Outpatient Appointment Scheduling given Individual Day-Dependent No-Show Predictions. European Journal of Operational Research, 240(1) 245-257.

  15. Samorani, M., M. Laguna. 2012. Data-Mining-Driven Neighborhood Search, INFORMS Journal on Computing, 24(2) 210-227.

  16. Samorani, M., M. Laguna, K.R. DeLisle, D. Weaver. 2011. A Randomized Exhaustive Propositionalization Approach for Molecule Classification. INFORMS Journal on Computing, 23(3) 331-345.

  17. Better, M, F. Glover, M. Samorani. 2010. Classification by Vertical and Cutting Multi-Hyperplane Decision Tree Induction. Decision Support Systems, 48(3) 430-436.

Major Conference Publications

  1. Samorani, M., S, Harris. 2019. The Impact of Probabilistic Classifiers on Appointment Scheduling with No-Shows. The 40th International Conference on Information Systems (ICIS). Munich, Germany. Historical acceptance rate ~17%, 1st ranked Information Systems conference.

  2. Ahmed, F., M. Samorani, C. Bellinger, O. Zaiane. 2016. Advantage of Integration in Big Data: Feature Generation in Multi-Relational Databases for Imbalanced Learning. IEEE International Conference on Big Data. 19% acceptance rate.

  3. Samorani, M., F. Ahmed, O. Zaiane. 2016. Automatic Generation of Relational Attributes: An Application to Product Returns. IEEE International Conference on Big Data. 19% acceptance rate.

  4. Samorani, M. 2015. Automatically Generate a Flat Mining Table with Dataconda. The International Conference of Data Mining (ICDM) Workshops.

  5. Samorani, M. 2015. Dataconda: A Software Framework for Mining Relational Databases. International Conference on Advances in Databases, Knowledge, and Data Applications (DBKDA), May 24 - 29, 2015 - Rome, Italy. Acceptance rate: 29%.

Other Refereed Publications

  1. Samorani, M., M. Laguna. 2017. Mining High-Quality Solutions to Learn Effective and Interpretable Heuristics. Proceedings of 2017 INFORMS Workshop on Data Mining and Analytics (DMA 2017). C. Iyigun, R. Moghaddess, M. Samorani, eds.

  2. Harris, S., M. Samorani. 2017. The Impact of No-Show Predictions on Appointment Scheduling. Proceedings of 2017 INFORMS Workshop on Data Mining and Analytics (DMA 2017). C. Iyigun, R. Moghaddess, M. Samorani, eds.

  3. Samorani, M. 2014. Automatic Generation of Relational Independent Variables. Proceedings of 2014 INFORMS Workshop on Data Mining and Analytics (DMA 2014). D. Sundaramoorthi, H. Yang, eds.

  4. Poursaeidi, M. H., M. Samorani, and O. E. Kundakcioglu. 2013. “Offshore Wind Farm Layout Optimization: What is the Hype?” in Proceedings of the Industrial and Systems Engineering Research Conference, ID: 952.

  5. Samorani, M. 2013. The wind farm layout optimization problem. In Handbook of Wind Power Systems (pp. 21-38). Springer Berlin Heidelberg.

  6. Samorani M., L. LaGanga. A Stochastic Programming Approach to Improve Overbooking in Clinic Appointment Scheduling, POMS Annual Meeting 2011.

Work in Progress

  1. Implicit Racial Bias In Healthcare Scheduling Delays (with Nan Liu, Shannon Harris, and Haibing Lu)

  2. The Role of Race and Socio-Economic Factors in Appointment No-Shows (with Shannon Harris and Paolo Roma)

  3. Racial Disparities in Direct Waiting Time in Appointment Scheduling Systems (with Shannon Harris and Karen Hicklin)