2016 INFORMS Annual Meeting Program

SC04

INFORMS Nashville – 2016

3 - Load Forecasting Using Support Vector Machine With Optimized Parameters Olufemi A. Omitaomu, Oak Ridge National Laboratory, omitaomuoa@ornl.gov Load forecasting is central to most of the energy transaction decisions in power systems planning and energy markets. Until now, most approaches for forecasting energy demand rely on monthly electrical consumption data. The emergence of smart meters is changing the data landscape for electric utility companies, and creating opportunities for utility companies to collect and analyze energy consumption data at a much finer temporal resolution. To enhance the estimation of energy demand at the household and network levels, we present an on-line accurate support vector regression technique that uses optimized regression parameters for forecasting real-time energy demand using smart meters data. 4 - Catch Me If You Can: Detecting Pickpocket Suspects From Large-scale Transit Records Chuanren Liu, Drexel University, chuanren.liu@drexel.edu Massive data collected by automated fare collection (AFC) systems provide opportunities for studying both personal traveling behaviors and collective mobility patterns in the urban area. We creatively leverage such data for identifying thieves in the public transit systems. We develop a thief active tracking system that identifies pickpocket suspects based on their daily transit records. We first extract a number of features from each passenger’s daily activities in the transit systems. Then, we exploit a combination of outlier detection and classification models to identify thieves, who exhibit abnormal traveling behaviors. SC02 101B-MCC Quality and Statistical Decision Making in Health Care Applications Sponsored: Data Mining Sponsored Session Chair: Cao Xiao, University of Washington, 3900 Northeast Stevens Way, MEB, Seattle, WA, 98195, United States, xiaoc@uw.edu Co-Chair: Shuai Huang, University of Washington, Seattle, WA, United States, shuaih@uw.edu 1 - Modeling And Analysis Of The Waiting Time Of Rapid Response Process In Acute Care Nan Chen, Tsinghua University, Room 615, Shunde Building, Tsinghua University, Haidian District, Beijing, 100084, China, chenn618@gmail.com, Xiaolei Xie, Li Zheng Improving the efficiency of rapid response process in acute care plays a significant role to ensure patient safety. We develop an analytical method to evaluate the waiting time and its variability. We discussed the structural properties and continuous improvement by adding care providers. A bottleneck indicator is introduced and a simple approximation formula is obtained. Case study is introduced to illustrated the application of the method. 2 - Modeling And Prediction Of The Mental Health Conditions Of Web Users Qingpeng Zhang, City University of Hong Kong, 1, Hong Kong, brianzqp@gmail.com The digital footprints of Web users left on the Web presents important proxies of their health conditions. In this research, we propose novel machine learning algorithms to model and predict the mental health conditions of Web users based on their online activities on social media. The preliminary results show the potential of using the open source social media data to infer the mental health conditions of people, and help health providers make better decisions. 3 - Learning Semantics Behind Health Status Disclosure On Twitter User generated content in social media is increasingly acknowledged as a rich resource for research into health problems. We in this talk present a framework to investigate how semantics are related with disclosure routines for 34 health issues. Our findings show that health issues related with family members, high medical cost and social support (e.g., Alzheimer’s Disease, cancer, and Down syndrome) lead to tweets that are more likely to disclose another individual’s health status, while tweets with more benign health issues (e.g., allergy, arthritis, and bronchitis) with biological processes (e.g., health and ingestion) and negative emotions are more likely to contain self-disclosures. Zhijun Yin, Vanderbilt University, Nashville, TN, 37203, United States, zhijun.yin@vanderbilt.edu, Bradley Malin

4 - Hospital Operational Health Monitoring: Enabling Organizational Communication Of Key Indicators And Analytics Diego A. Martinez, Scott R. Levin, Matthew F. Toerper, Johns Hopkins University School of Medicine, Baltimore, MD, dmart101@jhmi.edu Most hospitals have adopted electronic medical records, yet leveraging these data to optimize hospital operations remains a challenge. Grounded in human-com- puter interaction and visualization theory, we built a web app to facilitate data exploration and trend analysis. The app allows users to directly explore big data and scientifically assess whether or not an intervention is impacting hospital per- formance. Keeping clinicians and hospital leadership informed about practice operations can help align them with organizational goals, ultimately leading to better financial performance. SC03 101C-MCC Doing Good with Good OR I Invited Session Chair: Karen Smilowitz, Northwestern University, 2145 Sheridan Road RM D239, Evanston, IL, 60208, United States, ksmilowitz@northwestern.edu 1 - The Operational Challenges Of Sharing-Economies: An Optimal Re-balancing Mechanism For The Bike-Sharing Industry Pantelis Loupos, Department of Operations Management, Kellogg School of Management, Northwestern University, Evanston, IL 60208, Can Urguny Bike-sharing programs have been gathering momentum, but their expansion poses operational challenges. We propose a novel solution to the bike re-balanc- ing problem, that is centered around the actions of the riders instead of utilizing trucks for re-balancing. Our findings indicate great promise, whose adoption by bike sharing operators could have a positive impact on the industry. 2 - The Humanitarian Pickup And Distribution Problem Ohad Eisenhandler, Department of Industrial Engineering, Tel Aviv University, Tel Aviv, Israel, ohadeis@gmail.com, Michal Tzur We address the logistic challenges of food banks, which collect donated food from suppliers and distribute it to welfare agencies. We model the problem as a routing – resource allocation problem. Motivated by the activity of Israeli and American organizations, we introduce an innovative objective function, which balances equity and effectiveness in this operation, and propose exact and heuristic solution methods. 3 - Data Analytics For Optimal Detection Of Metastatic Prostate Cancer

Christine Barnett, Department of Industrial & Operations Engineering, University of Michigan, 1205 Beal Avenue, Ann Arbor, MI 48109, clbarnet@umich.edu, Selin Merdan

We used data-analytics approaches to develop, calibrate, and validate predictive models to help urologists make prostate cancer staging decisions. These models were used to design guidelines that weigh the benefits and harms of radiological imaging. The Michigan Urological Surgery Improvement Collaborative imple- mented these guidelines which miss less than 1% of metastatic cancers while reducing unnecessary imaging by more than 40%.

SC04 101D-MCC Gas-Power Market Integration Sponsored: Energy, Natural Res & the Environment, Energy I Electricity Sponsored Session

Chair: Robert Brooks, President, RBAC Inc, 14930 Ventura Blvd. Ste. 210, Sherman Oaks, CA, 91403, United States, rebrooks@rbac.com 1 - Analysis Of Gas / Electric Integration And Coordination In The Eastern Interconnection Of The United States And Canada Sara Wilmer, Levitan & Associates, Inc., sw@levitan.com Levitan & Associates has conducted recent analyses of gas-electric integration and coordination on behalf of the Eastern Interconnection Planning Collaborative and the Department of Energy. These analyses examined whether future electric sector demand for natural gas will be able to be accommodated by the available natural gas infrastructure as renewable penetration expands and coal-fired resources are retired. This case study will describe the modeling tools and integrated modeling framework used to conduct the work, and challenges faced both in the representation of real-world gas and electric systems in the selected modeling tools and in the integration of the different modeling tools.

69