INFORMS Nashville – 2016
18
SA02
SA02
101B-MCC
Healthcare Analytics and Medical Decision Making
Sponsored: Data Mining
Sponsored Session
Chair: Hasan Kartal, University of Massachusetts Lowell, One
University Avenue, Lowell, MA, 01854, United States,
Hasan_Kartal@uml.edu1 - PPMF: A Patient-based Predictive Modeling Framework For Early
ICU Mortality Prediction
Mohammad Amin Morid, The David Eccles School of Business,
University of Utah, 130 University Village, Salt Lake City, UT,
United States,
amin.morid@business.utah.edu,
Olivia R. Liu Sheng, Samir Abdelrahman
This paper presents a patient based predictive modeling framework (PPMF) to
improve the performance of early ICU mortality prediction. PPMF consists of
three main components. The first component captures dynamic changes of
patients’ status in the ICU using their time series data. The second component is a
local approximation algorithm that classifies patients based on their similarities.
The third component is a Gradient Decent wrapper that updates feature weights
according to the classification feedback. Experiments show that PPMF
significantly outperforms: (1) the severity score systems, (2) the aggregation based
classifiers, and (3) baseline feature selection methods.
2 - The Emergency Response Community Effectiveness Modeler:
A Simulation Modeling Tool To Analyze EMS vs.
Smartphone-based Samaritan Response
Michael Khalemsky, Graduate School of Business Administration,
Bar Ilan University, Ramat Gan, Israel,
khalemsky@gmail.com,David G. Schwartz
Smartphones and location-based social networking technologies present an
opportunity to re-engineer certain aspects of emergency medical response by
establishing Emergency Response Communities (ERC). The ERC Effectiveness
Modeler (ERCEM) estimates the efficacy of smartphone-based Samaritan
response for given medical condition and geographic region. The ERCEM uses
parameters such as population density, prescription adherence, smartphones
penetration etc. and performs Monte Carlo simulation to compare potential ERC
response to traditional EMS response. We present the modeler and show how it
assessed effectiveness of ERC for anaphylaxis in the USA based on data from the
NEMSIS project.
3 - Public Health Data Sharing With Privacy Protection
Hasan Kartal, Manning School of Business, University of
Massachusetts Lowell, Lowell, MA, 01850, United States,
hasan_kartal@uml.eduThis study examines privacy disclosure risks in health data when patients have
multiple records in a dataset. Existing data privacy approaches typically assume
that each individual in a dataset corresponds to a single record, which tends to
underestimate the disclosure risks in the multiple-record problems. We propose a
new privacy measure, called g-balance, and develop an efficient algorithm based
on the g-balance measure to protect against the multiple-record linkage attacks.
The effectiveness of the proposed approach is demonstrated in an experimental
study using real-world data.
SA03
101C-MCC
Nicholson Student Paper Prize I
Invited: Nicholson Student Paper Prize
Invited Session
Chair: Maria Esther Mayorga, North Carolina State University, 400
Daniels Hall, Raleigh, NC, 27695, United States,
memayorg@ncsu.edu1 - Nicholson Student Paper Prize
Maria Esther Mayorga, North Carolina State University,
Dept. of Industrial & Systems Engineering, Raleigh, NC, 27695,
United States,
memayorg@ncsu.eduThis session highlights the finalists for the 2016 George Nicholson Student Paper
Competition.
2 - A Necessary and Su cient Condition for Throughput Scalability of
Fork and Join Networks with Blocking
Yun Zeng, Ohio State University, Columbus, OH, United States,
zeng.153@buckeyemail.osu.edu, Augustin Chaintreau, Don
Towsley, Cathy Xia
3. Household-level Economics Of Scale In Transportation
Mehdi Behroozi, Northeastern University, Boston, MA, 02115,
United States,
behro040@umn.edu4. Online Decision-Making With High-Dimensional Covariates
Hamsa Bastani, Stanford University, Stanford, United States,
bayati@stanford.eduBig data has enabled decision-makers to tailor choices at the individual-level in a
variety of domains such as personalized medicine and online advertising. This
involves learning a model of decision rewards conditional on individual-specific
covariates. In many practical settings, these covariates are high-dimensional;
however, typically only a small subset of the observed features are predictive of a
decision’s success. We formulate this problem as a multi-armed bandit with high-
dimensional covariates, and present a new efficient bandit algorithm based on the
LASSO estimator. Our regret analysis establishes that our algorithm achieves
near- optimal performance in comparison to an oracle that knows all the problem
parameters. The key step in our analysis is proving a new oracle inequality that
guarantees the convergence of the LASSO estimator despite the non-i.i.d. data
induced by the bandit policy. Furthermore, we illustrate the practical relevance of
our algorithm by evaluating it on a real-world clinical problem of warfarin dosing.
5 - Distributionally Robust Stochastic Optimization With Wasserstein
Distance
Rui Gao, Georgia Institute of Technology, Atlanta, GA, United
States,
rgao32@gmail.comSA04
101D-MCC
Electricity Markets and Contract Design
Sponsored: Energy, Natural Res & the Environment,
Energy I Electricity
Sponsored Session
Chair: Edward James Anderson, University of Sydney,
H70 - Abercrombie Building, Sydney, NSW 2006, Australia,
edward.anderson@sydney.edu.au1 - Retail Equilibrium With Switching Consumers In
Electricity Markets
Carlos Ruiz Mora, Universidad Carlos III de Madrid, Madrid, Spain,
caruizm@est-econ.uc3m.es, F. Javier Nogales, F. Javier Prieto
We consider a game theoretical model where asymmetric retailers compete in
prices to increase their profits by accounting for the utility function of switching
consumers. Consumer preferences for retailers are uncertain and distributed
within a Hotelling line. We analytically characterize the equilibrium of a retailer
duopoly, establishing its existence and uniqueness conditions for a wide class of
utility functions. The duopoly model is extended to a multiple retailer case.
2 - Flow-based Market Coupling In The European Electricity Market
Mette Bjørndal, Professor, NHH Norwegian School of Economics,
Bergen, Norway,
Mette.Bjorndal@nhh.noFrom May 2015, the Flow-Based Market Coupling (FBMC) model replaced the
Available Transfer Capacity (ATC) model in parts of the European power market.
The FBMC model aims to enhance market integration and to better monitor the
physical power flow, and it is expected to lead to increased social welfare in the
day-ahead market and more frequent price convergence between different
market zones. This paper gives a discussion on mathematical formulations of the
FBMC model and the procedures of market clearing. We examine the FBMC
model in two test systems and show the difficulties of implementing the model in
practice.
3 - Negotiating Forward Contracts With Private Information
Edward James Anderson, University of Sydney,
edward.anderson@sydney.edu.auWe consider the use of forward contracts to reduce risk for firms operating in a
spot market. Firms have private information on the distribution of prices in the
spot market. We discuss different ways in which firms may agree on a forward
contract (offers to a broker and direct bargaining). We also discuss an equilibrium
in which two firms each offer a supply function and the clearing price and
quantity for the forward contracts are determined from the intersection. In this
context a firm can use the offer of the other player to augment its own
information about the future price. It is interesting that these sophisticated
strategies are likely to produce worse outcomes for both firms.