Informs 2015

465 / 552

INFORMS Philadelphia – 2015
463
WD30
30-Room 407, Marriott
Information Systems III
Contributed Session
Chair: Miaomiao Lu, Huazhong University of Science and Technology,
Room403, South Student Hostel, the Huazh, Wuhan, China,
1 - Online Social Networks: The Social Influence of Sentiment
Content on Digital Product Diffusion
Tung Cu, Louisiana State University, 2200 Business Education
Complex, Nicholson Extention, Baton Rouge, LA, 70803,
United States of America, Helmut Schneider,
James Van Scott
The study explores the role of user-generated content (UGC) during the diffusion
process of digital artifacts. Data collection is conducted on 260 new digital
products and more than 105 thousand social network nodes. The overall finding
shows that Volume of Post and UGC Sentiment have a dynamic impact on
diffusion of digital products. But, the relationships among them depend on certain
situations. The study sheds light on the crowding power and the long-tail effect in
online social networks.
2 - How Much to Open, How Fast to Fix? Effects of Making the
Software Open Source
Rakesh Mallipeddi, PhD Student, Texas A&M University, 320
Wehner, 4217 TAMU, College Station, TX, 77843-4217, United
States of America, Subodha Kumar,
Ram Gopal, Emre Demirezen
We develop empirical and analytical models to examine the effects of making
software open on the overall quality of software systems and behavior of software
vendor. We derive and develop optimal strategies for software vendor to allocate
resources for maintenance of existing software while developing new software.
3 - Towards a Theoretical Framework of IT-enabled
Operations Strategy
Yeming Gong, Associate Professor, EMLYON Business School,
12 Rue Dunoir, lyon, France, Hongyi Mao,
Ryad Titah, Oliver Yao
By an integrated analysis of quantitative data from more than 100 organizations
in Europe, Asia and North America and qualitative data from 56 cases, this paper
presents a theoretical framework of IT-enabled operations strategy with the
objective of investigating “How does information technology leverage resources
and processes for operational agility?”
4 - Timing, Diffusion,and Substitution of Generations of
Technological Innovations
Miaomiao Lu, Huazhong University of Science and Technology,
Room403, South Student Hostel, The Huazh, Wuhan, China,
Diffusion processes across generations and over time have become increasingly
complex and multifaceted in recent years. We discuss efforts to model
simultaneously the substitution of successive generation of a durable
technological innovation,and the diffusion of the technology.Empirical and
normative implications of the proposed model are explored for four generations
on Microsoft Windows operating system:win Vista;win xp;win 7;win 8.
WD31
31-Room 408, Marriott
Data Mining in Medical and Sociological
Decision Making
Sponsor: Data Mining
Sponsored Session
Chair: Chitta Ranjan, Georgia Institute of Technology,
Ferst Drive NW, Atlanta, GA, United States of America,
Co-Chair: Kamran Paynabar, School of Industrial and Systems
Engineering, 755 Ferst Drive, NW, Atlanta, GA, 30332,
United States of America,
1 - Online, Semi-Parametric Estimation of Treatment Allocations for
the Control of Emerging Epidemics
Eric Laber, 211 Devonhall Lane, Cary, NC, 27518,
United States of America,
A key component in controlling the spread of an epidemic across a network of
individuals is deciding where, when, and to whom to apply an intervention. An
allocation strategy formalizes this process as a sequence of functions that map up-
to-date information on the epidemic to a subset of nodes targeted for treatment.
We derive estimating equations for the optimal allocation strategy that do not
require a model the system dynamics and that scale to very large problems.
2 - A Novel Sequence Kernel Graph Transform for Clustering
and Visualization
Chitta Ranjan, Georgia Institute of Technology,
755 Ferst Drive NW, Atlanta, GA, United States of America,
Samaneh Ebrahimi,
Kamran Paynabar
We propose a novel sequence kernel graph (SKG) transform for non-parametric
feature extraction on sequence data. The proposed method is accurate, faster than
existing methods and parallelizable. The SKG transform can be used for finding
similarity between sequences, and hence, alignment-free clustering. It can be
extended to perform bi-clustering, and graph visualization of sequences; with
application on various behavioral data (clickstream, purchase pattern), protein
and gene sequences, etc.
3 - A Transfer Learning Approach for Predictive Modeling of
Degenerate Biological Systems
Jing Li, Arizona State University, Tempe, AZ,
United States of America, Na Zou
Transfer learning, as a statistical modeling approach, refers to methods that
integrate knowledge of old domains and data of a new domain, in order to
develop a model for the new domain that is better than using the data of the new
domain alone. We propose a transfer learning method for predictive modeling
and apply it to degenerate biological systems. Theoretical results and findings
from real-data analysis will also be presented.
4 - Bayesian Learning Without Recall: A Naive Social Learning Model
Mohammad Amin Rahimian, Graduate Research Fellow,
University of Pennsylvania, Levine 4F, University of
Pennsylvania, 3330 Walnut Street, Philadelphia, PA, 19104,
United States of America, , Ali Jadbabaie
We analyze a model of learning and belief formation in networks in which agents
follow Bayes rule yet they do not recall their history of past observations and
cannot reason about how other agents are making their decisions. This model
avoids the complexities of fully rational inference and also provides a behavioral
foundation for non-Bayesian updating. We present the implications of the choice
of signal and action structures for such agents leading to familiar update forms.
WD32
32-Room 409, Marriott
Data Mining and Optimization
Contributed Session
Chair: Fatma Yerlikaya Ozkurt, Middle East Technical University,
Institute of Applied Mathematics, Ankara, Turkey,
1 - Data Classification via Cluster Covering
Zhengyu Ma, Korea University, Room 551,Engineering Building,
Seoul, 136-713, Korea, Republic of,
Kwangsoo Kim, Hong Ryoo
For data classification, a homogeneous cluster containing only one type of data
can easily be identified via neighborhood measure. Using clusters for patterns in
LAD, one can obtain a set of homogeneous clusters and next optimize their
interplay to discover a decision theory. This new framework of supervised data
analytics inherits the major advantage of LAD while it avoids redundant data
binarization and also the difficult stages of support feature selection and pattern
generation in LAD.
2 - Wastewater Sewerage Treatment Plant Aeration Process
Optimization: A Data-driven Approach
Anoop Verma, Research Associate, Wayne State University,
4815 4th St, Detroit, MI, 48201, United States of America,
, Kai Yang, Ali Asadi
Being water quality oriented, large-scale industries such as wastewater treatment
plants tend to overlook potential savings in energy consumption. Wastewater
treatment process includes energy intensive equipment such as pumps and
blowers to move and treat wastewater. Presently, a data-driven approach has
been applied for aeration process modeling and optimization of one large scale
wastewater in Midwest. A great deal of saving in energy can be made while
keeping the water quality within limit.
3 - Optimal Experimental System Design
Alireza Mohseni, Oregon State University, 204 Rogers,
Corvallis, OR, 97331, United States of America,
, David Kim
This research examines how the design of a factorial experiment can be modeled
as a cost and time constrained discrete optimization problem. Initial results with
respect to creating a statistical model for an experiment will be presented.
WD32