Michael T. Lash

Michael T. Lash

♦ Assistant Professor at the University of Kansas School of Busines ♦ Technology, Operations, and AI Consultant

Posts by Collection

portfolio

publications

Early prediction of movie success-what, who, and when

Published in Proceedings of the 2015 International Conference on Social Computing, Behavioral-Cultural Modeling, & Prediction (SBP), 2015

This paper is a preliminary study investigating the drivers of movie success from a pre-production disposition.

Recommended citation: M.T. Lash, S. Fu, S. Wang and K. Zhao, Early prediction of movie success-What, who, and when, in Proceedings of the 2015 International Conference on Social Computing, Behavioral-Cultural Modeling, & Prediction (SBP), 2015. http://michael-lash.github.io/files/sbp-2015.pdf

Early predictions of movie success: The who, what, and when of profitability

Published in Journal of Management Information Systems (JMIS), 2016

This paper investigates the predictability of movie success (profit) utilizing features (variables) strictly available during the pre-production stage. We investigate the features that are important to the prediction problem and propose a "Movie Investor Assurance System" (MIAS).

Recommended citation: M.T. Lash and K. Zhao, Early predictions of movie success: The who, what, and when of profitability, Journal of Management Information Systems (JMIS), 33(3):874-903, 2016. https://www.tandfonline.com/doi/abs/10.1080/07421222.2016.1243969

Generalized Inverse Classification

Published in Proceedings of the 2017 SIAM International Conference on Data Mining (SDM), 2017

We propose an updated optimization methodology to generalize the inverse classification involving only very minor assupmtions. The proposed method and framework can be used with virtually any classifier.

Recommended citation: M.T. Lash, Q. Lin, W.N. Street, J.G. Robinson and J. Ohlmann, Generalized Inverse Classification, in Proceedings of the 2017 SIAM International Conference on Data Mining (SDM), pp. 162-170, 2017. http://michael-lash.github.io/files/lash_sdm_2017.pdf

A web-based registry for patients with sarcoidosis

Published in Sarcoidosis, vasculitis, and diffuse lung diseases, 2017

This work proposes a data repository for patients suffering from the disease, sarcoidosis. The data is subsequently analyzed to tease out insights.

Recommended citation: A.K. Gerke, F. Tang, M.T. Lash, J. Schappet, E. Phillips and P.M. Polgreen, A web-based registry for patients with sarcoidosis, Sarcoidosis vasculitis and diffuse lung diseases (SVDLD), 34(1):26-34, 2017. http://michael-lash.github.io/files/sarcoid_paper.pdf

A Large-Scale Exploration of Factors Affecting Hand Hygiene Compliance Using Linear Predictive Models

Published in 2017 International Conference on ealthcare Informatics (ICHI), 2017

We examine the factors that influence hand hygiene compliance in hospitals utilizing real-world hand hygiene sensor data from 19 different facilities.

Recommended citation: M.T. Lash, J. Slater, P.M. Polgreen, and A.M. Segre, A Large-Scale Exploration of Factors Affecting Hand Hygiene Compliance Using Linear Predictive Models, in Healthcare Informatics (ICHI), 2017 International Conference on, pp. 66-73, 2017. http://michael-lash.github.io/files/hh_linear_models_ichi_2017_pub.pdf

A budget-constrained inverse classification framework for smooth classifiers

Published in 2017 IEEE International Conference on Data Mining Workshops (ICDMW), 2017

We propose a method and framework for the inverse classification problem that assumes the model in question is differentiable with an L-Lipschitz continuous gradient.

Recommended citation: M.T. Lash, Q. Lin, W.N. Street and J.G. Robinson, A budget-constrained inverse classification framework for smooth classifiers, in Data Mining Workshops (ICDMW), 2017 IEEE International Conference on, pp. 1184-1193, 2017. http://michael-lash.github.io/files/budget_constrained_icdmw2017.pdf

Learning Rich Geographical Representations: Predicting Colorectal Cancer Survival in the State of Iowa

Published in 2017 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), 2017

In this work, we propose a method to learn latent representations from spatial (georaphic) data to predict colorectal cancer survival curves, specifically focusing on data covering the US state of Iowa.

Recommended citation: M.T. Lash, Y. Sun, X. Zhou, C.F. Lynch, and W.N. Street, Learning Rich Geographical Representations: Predicting Colorectal Cancer Survival in the State of Iowa, in Bioinformatics and Biomedicine (BIBM), 2017 IEEE International Conference on, pp. 778-785, 2017. http://michael-lash.github.io/files/rich_geo_reps_bibm_2017.pdf

Deriving Enhanced Geographical Representations via Similarity-based Spectral Analysis: Predicting Colorectal Cancer Survival Curves in Iowa

Published in International Journal of Data Mining and Bioinformatics (IJDMB), 2019

This work extends the method and analysis of our 2017 BIBM paper; we propose a method to learn from geographical data to predict colorectal cancer survival curves for patients in the US state of Iowa.

Recommended citation: M.T. Lash, M. Zhang, X. Zhou, C.F. Lynch, and W.N. Street, Deriving Enhanced Geographical Representations via Similarity-based Spectral Analysis: Predicting Colorectal Cancer SurvivalCurves in Iowa, International Journal of Data Mining and Bioinformatics (IJDMB), 21(3):183-211, 2018. http://michael-lash.github.io/files/ijdmb-2019.pdf

21 Million Opportunities: A 19 Facility Investigation of Factors Affecting Hand Hygiene Compliance via Linear Predictive Models

Published in Journal of Healthcare Informatics Research (JHIR), 2019

This paper extends the analysis of our 2017 ICHI paper investigating health care worker hand hygiene compliance.

Recommended citation: M.T. Lash, J. Slater, P.M. Polgreen, and A.M. Segre, 21 Million Opportunities: A 19 Facility Investigation of Factors Affecting Hand Hygiene Compliance via Linear Predictive Models, Journal of Healthcare Informatics Research (JHIR), 3(4):393-413, 2019. http://michael-lash.github.io/files/21_million_opportunities.pdf

Personalized Cardiovascular Disease Risk Mitigation via Longitudinal Inverse Classification

Published in 2020 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), 2020

We propose an inverse classification framework for longitudinal data to take advantage of an instance's progression over time.

Recommended citation: M.T. Lash and W.N. Street, Personalized Cardiovascular Disease Risk Mitigation via Longitudinal Inverse Classification, in Bioinformatics and Biomedicine (BIBM), 2020 IEEE International Conference on, pp. 2610-2617, 2020. http://michael-lash.github.io/files/personalized_cvd_ic_bibm2020.pdf

Optimal Sepsis Patient Treatment using Human-in-the-loop Artificial Intelligence

Published in Expert Systems with Applications (ESWA), 2021

In this paper we propose a method for optimizing IV fluid dosages for septic patients in the ICU.

Recommended citation: A. Gupta, M.T. Lash, S.K. Nachimuthu, Optimal Sepsis Patient Treatment using Human-in-the-oop Artificial Intelligence, Expert Systems with Applications (ESWA), 169:1-14, 2021. http://michael-lash.github.io/files/optimal_sepsis_patient_treatment.pdf

Predicting mobility using limited data during early stages of a pandemic

Published in Journal of Business Research (JBR), 2023

We examine the factors driving retail mobility in disparate geographical locations across the USA during the COVID-19 pandemic using linear predictive models; we propose a novel method to conduct this analysis.

Recommended citation: M.T. Lash, S. Sajeesh, O.M. Araz, Predicting mobility using limited data during early stages of a pandemic, Journal of Business Research (JBR), 157, 2023. http://michael-lash.github.io/files/jbr-2023.pdf

talks

teaching

BSAN 745: Advanced Machine Learning

Graduate Course, University of Kansas School of Business, Semesters: Spring 2021 - Spring 2023

This course takes a data-centric approach to machine learning, focusing on nontraditional data types, particularly text, networks/graphs, and images. These data types have become ubiquitous, invaluable sources of information and insight spanning the spectrum of industries and enterprise. It is therefore imperative that business analysts, data scientists, and machine learning practitioners have competency in not only dealing with these sources of data but in gleaning the treasure-trove of insights such data provide. This course equips students with the skills necessary to acquire, process, and learn from text, networks, and images, building on the foundations learned in the introduction to machine learning course (recommended, but not required). The primary emphasis in this course is text data; select lectures will be devoted to working with networks and images.

BSAN 995: Machine Learning Seminar

Graduate Course (PhD), University of Kansas School of Business, Semesters: Spring 2020, Spring 2023

This seminar introduces graduate students to the vast, exciting, and ever-growing research area of machine learning. Both historical and current research papers on the subject will be presented and discussed to cultivate a trajectory-based perspective of the field. Furthermore, the course adopts a “model-based machine learning” (C. Bishop) perspective and emphasis: the application should inform the selection and development of a particular machine learning model (i.e., is application-focused). Both supervised (learning from data with a target outcome) and unsupervised methods (learning from data without a definitive outcome) will be presented and discussed. The particular methods and applications covered are based on student interests, determined at the onset of the semester. The instructor introduces the field during the first few weeks of the semester, with instructor, guest, and student research paper presentations taking place thereafter. Student evaluation is based on paper presentations and a semester-long project.

BSAN/IST 326: Database Management Systems

Undergraduate Course, University of Kansas School of Business, Semesters: Spring 2020 - Spring 2023

This undergraduate course, offered to business analytics and information systems students, covers the database development process. The class begins with an overview of databases and the benefits of adopting such an approach. We then work through the database development process, beginning with requirementa analysis. The second half of the class focuses on SQL, beginning with data definition language (DDL) commands to create the tables in a database system. We then cover data manipulation language (DML) commands to add data to the tables. The last month or so of class focuses specifically on the SELECT command to derive insights directly from a database. We also briefly cover data control language (DCL) commands. The class uses Oracle cloud-based database management software and freely available cloud-based diagramming software.