2015-2016 department newsletter

Annual Report 2015-2016

WOOSTER MATH & COMPUTER SCIENCE

MCM/ICM Math Modeling Contest -ADVISED BY BOB WOOSTER AND MATT MOYNIHAN

Back in the cold, dark days of late January, The College of Wooster had four teams of math and computer science majors compete in a 96-hour, worldwide math modeling competition. In the contest each team can choose to do an MCM (Mathematical Contest in Modeling) or an ICM (Interdisciplinary Contest in Modeling) problem. In the MCM: 7421 teams participated. In the ICML: 5025 teams participated.

The teams from Wooster: • 1. Shiwani Varal ('18), Paige McKean ('18), and Haidar Esseili ('19): Our all-first-time-competitors team tackled ICM problem F and earned a Meritorious which put them in the top 20% of the 5025 teams! ICM problem F had to do with managing the current refugee and immigrant crisis in the Middle East and Europe. • 2. Khoa Nguyen ('18), Alex Iudice ('17), and Jacob Priest ('16) chose MCM problem C and earned an Honorable Mention rank. MCM problem C dealt with creating an optimal strategy to invest grant money in education. • 3. Robin Morillo ('17), Preston Pozderac ('17), and Alishan Premani ('18) took on MCM problem A and earned a Successful Participant rank. MCM problem A had to do with modeling the water temperature in a bathtub, and ways to keep the temperature uniform and constant as efficiently as possible. • 4. Ming Zhao ('17), Jianqiu Bai ('18), and Jinzhe Liu ('18) took on ICM problem E and earned an Honorable Mention rank. ICM problem E asked teams to model fresh water availability on the planet and develop methods to anticipate and combat water scarcity. These four teams worked very hard, did some very excellent modeling, and wrote some very nice papers. All of them did an outstanding job representing The College of Wooster on this national stage.

�1


SENIOR INDEPENDENT STUDY

�2

Michael Bush Math & Physics double major

-ADVISED BY JENNIFER BOWEN (MATH) AND CODY LEARY (PHYSICS)

A COMPARATIVE ANALYSIS OF PHOTON AND ELECTRON WAVE FUNCTIONS IN SPHERICALLY SYMMETRIC POTENTIALS

We derived the wave function for light in a finite spherical well potential by drawing an analogy to an electron in an analogous potential energy well. The theory behind vector calculus on a spherical basis was examined. The wave equation for light, derived from Maxwell’s equations, and the Pauli equation for electrons were combined into a unified form that was solved by separation of variables, infinite series solutions, and numerical methods. The potential well for light was established by considering an environment with a constant index of refraction inside a spherical boundary and a different, but still constant, index of refraction outside the boundary. It was determined that the radial part of the wave function oscillated more inside the boundary as radial quantum number increased. The distance between the origin and the first peak amplitude inside the boundary increased as the angular momentum quantum number increased. For light, the wave number was found to be complex, therefore the temporal part of the wave function was a dampened oscillator. The time constant of the temporal part of the wave function increased as angular momentum quantum increased and decreased when the radial quantum number increased. (Michael will attend the University of Delaware for graduate study in mathematics.)

Thomas Ames Math & Philosophy double major

-ADVISED BY PAM PIERCE (MATH) AND RON HUSTWIT (PHILOSOPHY)

LOGICISM AND THE CONTINUUM HYPOTHESIS

This inquiry into the philosophy of mathematics considers the undecidability of the continuum hypothesis with ZFC set theory as a unique case of mathematical knowledge for the logicist theory of mathematics. The first two chapters are dedicated to building the mathematical foundation needed to understand the generalized continuum hypothesis and grasp the essence of the proofs of Kurt Gödel and Paul Cohen, which together establish the undecidability of the continuum hypothesis within the axioms of ZFC set theory. The final chapter presents both Gottlob Frege and Bertrand Russell’s brand of logicism and analyzes how well their theories are able to account for such a unique conjecture of mathematics. The analysis in the final chapter first establishes which criteria determine if a theory can be said to account for the continuum hypothesis. Informed by such criteria, the analysis concludes with an argument that neither Gottlob Frege nor Bertrand Russell’s logicist theory can adequately account for the continuum hypothesis.


Nakesha Newsome Math major

-ADVISED BY JENNIFER BOWEN

DIRICHLET TESSELATIONS

A Dirichlet domain is an array of points with a single source point located in a region. This region is defined such that every point within that region is closest to the source point of that territory than to any other source point. The collection of these regions will tessellate the plane forming Dirichlet tessellations. These tessellations have many applications in a multitude of different fields, making them very useful. We can recognize a given tessellation as Dirichlet by investigating the boundary lines, examining the sources within each region in relation to other sources and in several other ways. These ideas form the basic theorems and lemmas that allow us to determine if a tessellation is Dirichlet. This work will explore those theorems and lemmas as well as the definition of Dirichlet tessellation. Then we will examine an algorithm that will allow us to compute a Dirichlet tessellation using nearest neighbor techniques from computer science.

�3

Paroma Palchoudhuri Math & Physics double major

-ADVISED BY JOHN RAMSAY (MATH) AND SUSAN LEHMAN (PHYSICS)

AN EXPERIMENTAL AND MATHEMATICAL STUDY OF AVALANCHE BEHAVIOR

The avalanche distribution for a combined effect of tuning parameters was studied in a slowly-driven critical conical pile of steel beads. The tuning parameters were specifically drop height and cohesive forces between the beads. The cohesion was generated by placing the pile in between a pair of Helmholtz coils that created a uniform magnetic field in which the pile rested. The avalanche distribution was studied for a drop height of 2 cm at six different cohesion levels created by coil currents of 0, 300, 500, 630, 750 and 900 mA. Fractional occurrences of avalanches were calculated based on their size; the probability of large avalanches increased and that of mid-sized avalanches decreased with increasing cohesion resulting in humps in the probability distributions. Excessive number of double bead drops significantly effect the probability distribution of avalanches increasing the probability of mid-sized avalanches and decreasing that of smaller avalanches. New analysis techniques were developed and explored in order to study the inter-event time distributions and avalanche behavior. The waiting time after an avalanche in conjunction with k-means clustering was used to find the avalanche size threshold separating local from non-local avalanches, which was found to be at approximately 70 beads. For the inter-event time probability densities, it was found that the system spanning avalanche regime was well fit to a Brownian passage-time distribution for very large avalanches only, while the smaller and mid-sized avalanche regime was better fit by the Weibull distribution. (Popi has accepted a position as data scientist/software developer at Embark Corporation in New York City.)


Jeffrey Giegold Math major

-ADVISED BY ROBERT WOOSTER

SHOOTERS VS GOALIES: ANALYZING SHOOTOUTS IN THE NHL WITH GAME THEORY

Game theory is a relatively new field of mathematics that explores analysis of strategic decision-making in competitive situations. Game theory primarily focuses on finding optimal actions for anyone involved, and these actions present themselves as probability distributions over many different possible simple actions. It can be applied to many different fields of study, but this research applies it strictly to shootouts in the National Hockey League, a typical one-on-one competitive situation that Game Theory can be used to analyze. This study uses the ideas of The Minimax Theorem and The Equilibrium Theorem to build feasible games that can represent shootouts and compares the results of this construction to what we would expect from real life situations.

Michael Sokolich Math and Computer Science double major

-ADVISED BY MATTHEW MOYNIHAN (MATH) AND RHYS PRICE JONES (COMPUTER SCIENCE)

AN INVESTIGATION OF THE BURNING NUMBERS OF GRAPHS AND TREES

The burning number of a graph is a graph parameter that models the spread of a contagion in a system. It is based on graph searching problems such as the firefighter problem and graph domination. The smaller the burning number of a graph, the faster the contagion spreads. In this paper we develop a program to calculate the burning number of a tree. We use this to investigate the correlation between tree properties and the burning number of a tree. We conclude with an alternative program design, and discuss open problems related to the burning number of trees and connected graphs.

�4


Rob Handloser Math major

-ADVISED BY JAMES HARTMAN

HALL OF FAME CLASSIFICATION USING LOGISTIC REGRESSION ANALYSIS

Election in the Naismith Memorial Basketball Hall of Fame is the crowning achievement of any basketball player’s career. This exclusive group, which includes players, coaches, referees, teams, and other contributors to the game, contains just over 300 inductees. We construct several predictive models that classify a given player as either being elected into the Hall of Fame, or not, based on their career statistical output and achievements. The goal here is to make future predictions about who might be included into the Hall of Fame. As a basis for this research, we provide an understanding of (binary) logistic regression theory. This regression method also leads us to areas in mathematics such as maximum likelihood estimation and a multivariable Newton-Raphson iteration method. This independent study project touches on the histories of the game of basketball, what it takes to become enshrined into the Naismith Memorial Basketball Hall of Fame, and of logistic regression’s origin. Furthermore, we use maximum likelihood theory, along with the Newton-Raphson iteration method, to help us understand how logistic regression works. Finally, we experiment with a few different models, explore diagnostics for those models, and elaborate on their implications.

Kinsey Small Computer Science major

-ADVISED BY DENISE BYRNES AND RHYS PRICE JONES

JUDGING A BOOK BY ITS COVER: WEB AND MOBILE INTERFACE DESIGN

Usability is a component of software development that is often overlooked, and even when considered, is often considered incorrectly. This independent study analyzes the user interface guidelines of three usability pioneers in relation to web and mobile interface design, uses their guidelines in the development of the front-end design of the Computer Science Department web site on the CS Web server, and analyzes the resulting website in terms of its adherence to those guiding principles.

�5


Michael Bonadio Computer Science major

-ADVISED BY DENISE BYRNES

DYNAMICALLY ADJUSTING VIDEO GAME DIFFICULTY USING MACHINE LEARNING TECHNIQUES

Balancing video game difficulty is a constant struggle to keeping the game challenging to maintain player interest but not overly challenging as to frustrate the player. While the player progresses through the game, their perception of too hard and too easy changes. Games often use a static difficulty curve to increase the difficulty as the player progresses through the game. However, dynamically adjusting the game difficulty creates a better fitting difficulty balancing, leading to a better player experience. To create such an experience, this work focuses on matching learning techniques decision trees and Fuzzy Inference Systems for the use of Dynamic Difficulty Adjustment. A decision tree learns how to evaluate a player’s performance based on in-game statistics and outputs a ranking. This ranking is then used by the Fuzzy Inference System, along with the player’s current health and damage output, to decide if the difficulty is increased, decreased, or left the same and to what degree it is changed. Therefore, a combination of decision trees and Fuzzy Inference Systems leads to a Dynamic Difficulty Adjustment.

Kiera Dobbs Math major

-ADVISED BY MATTHEW MOYNIHAN

OPTIMAL PLACEMENT OF FAMILY PLANNING CENTERS

This project investigates and begins to solve the problem of access to family planning services in the United States. We research and implement methods in Operations Research to optimize the location of publicly funded family planning centers in the U.S. by minimizing travel distance. The solution begins with a designated number of family planning centers for the country. An apportionment integer programming algorithm is then exercised to allocate centers to all the states based on population, percent of population in poverty, and state square mileage. At the state level, we use apportionment again to distribute centers to counties. At the county level, a facilities location integer programming model decides which county census tracts are assigned a center. This is done in a way that minimizes travel distance from tracts to centers. Here the travel distance data is weighted based on populations, to ensure that more people travel shorter distances. Finally, we use a set-covering integer programming algorithm to compute the placement and minimum number of abortion clinics required to satisfy a set maximum travel distance for a particular state. (Kiera will enter a Master’s program in Operations Research at Miami University of Ohio.)

�6


Lucas Hunt Math major

-ADVISED BY MATTHEW MOYNIHAN

THE EXTRA POINT RULE CHANGE AND ITS EFFECTS ON DECISION-MAKING FOR COACHES IN THE NFL The NFL has implemented a new rule change before the start of the 2015 season, moving the distance of the extra point back 13 yards to the 15-yard line. In an attempt to create more excitement with every play, the NFL rules committee feels strongly about its decision to adjust the level of difficulty for an extra point try. Consequently, the increased distance may negatively affect the overall success rates for extra points, causing the decision process for post-touchdown conversions to be more stressful for coaches. Thus, with the use of success rates, as well as expected point, there may be other external factors that potentially affect the decision between kicking the extra point and going for the two-point conversions.

Katelyn French Math major


BEES IN THE TRAP: A MATHEMATICAL APPROACH TO MODELING THE MUTUALISM BETWEEN FLOWERING PLANTS AND HONEY BEES

The mysterious disappearance of honey bee populations has caused worldwide concern in the past decade. This loss of honey bees not only affects beekeepers, but agriculture on a large scale, as honey bees are responsible for the pollination of many crops. The most recent research has suggested that pesticides and outside stressors may be the cause of what has been called colony collapse disorder (CCD). In this study, we use differential equation models and agent-based models to describe the mutualistic relationship between honey bees and the flowers they pollinate with a focus on understanding how external pressures can cause populations to collapse. This not only provides insight into CCD but also adds to the literature of models of interacting species, in which mutualism have been less of a focus. (Katelyn has accepted a position at Epic Systems in Madison, Wisconsin.)

�7


Adam Coppock Math major


APPLYING MODERN PORTFOLIO THEORY TO OPTIMAL PORTFOLIO CONSTRUCTION

Henry Markowitz discovered Modern Portfolio Theory while completing his doctorate thesis at the University of Chicago, with the goal of finding the optimal investment portfolio in mind. Under Markowitz’s work, a portfolio, or collection of securities, is studied using a combination of concepts from portfolio theory. These concepts include expected value, variance, and covariance. They combine to produce an expected return and risk for each attainable portfolio, which are then compared with one another to create an efficient frontier. The portfolios that live on the efficient frontier produce the most return for a given level of risk and are determined most desirable for portfolio managers because it returns the most bang for their buck. (Adam has accepted a position at Prudential Advisors in Pittsburgh.)

Jacob Priest Computer Science major

-ADVISED BY SOFIA VISA

ROBOTIC NAVIGATION USING IMAGE ANALYSIS

The main focus of this independent study is robotic navigation within a maze. We look at Artificial Intelligence, robotic navigation, and image analysis to show how far robotics has come, especially in the last twenty years when hardware and software became both more capable and inexpensive. We use image recognition software, known as OpenCV (Open Computer Vision) and a Raspberry Pi mini-computer to provide the robot the capability of analyzing images and performing various functions to manipulate images to each experiment’s needs. This independent study covers five experiments testing the capabilities of both Open CV, Raspberry Pi and their various functions and practicalities.

�8


Jacob London Computer Science major


SHOW ME WHAT I WANT: COMPLEXITIES IN CREATING AN EFFECTIVE USER INTERFACE

User interfaces are everywhere, they dictate how humans are able to interact with machines and their effectiveness determines the effectiveness of the user. This project looks at the underlying principles associated with creating an effective user interface and applies these guideline to the creation of a user interface in a video game demonstration level. The interface is designed using Unreal Engine 4 and, along with Battlefield 4, a currently popular video game, is subject to a usability test with four testers of varying degrees of experience. This project also looks at the power software suites – video game engines, describes the design and implementation process of creating the user interface, as well as the usability tests, and discusses the results and implications of those tests. The reader will understand the principles that determine an effective user interface, the process for which usability testing is conducted, why Unreal Engine 4 was chosen for this project, and the difficulties associated with creating a user interface for different types of users.

Catherine Eckbold Computer Science major

-ADVISED BY SOFIA VISA ARE YOU SURE YOU’RE IN THE RIGHT CLASSROOM? RESEARCHING THE LACK OF WOMEN IN COMPUTER SCIENCE WORKSHOP AND WEBSITE CREATION DESIGNED TO ENCOURAGE WOMEN IN COMPUTER SCIENCE

This I.S. project examines the fundamental disincentivizing reasons for the decline of women in the field of computer science. This thesis addresses this problem in three ways: (1) researching the reasons why women’s participation in the field of computer science is declining and presenting several ideas to address these reasons; (2) creating a workshop for middle school aged girls utilizing the research from (1), with an aim to spark and then maintain their interest in computer science; and (3) building a website to circulate the information gathered from (1) and (2) and to model and illustrate important data, facts and information pertinent to women in computer science.

�9


Sayantan Mitra Math and Economics double major

-ADVISED BY JAMES HARTMAN (MATH) AND AMYAZ MOLEDINA (ECONOMICS)

HOLDING HANDS AGAINST THE UNKNOWN: USING MARKOV CHAINS TO MODEL INFORMAL INSURANCE ARRANGEMENTS IN DEVELOPING SOCIETIES

In developing societies, in the absence of formal insurance markets, when a household faces a catastrophic financial shock, are they left completely on their own? In this I.S., I argue that these households have an informal insurance arrangement amongst themselves in which “current generosity is justified by future reciprocity” (Coate & Ravallion, 1993). Assuming that the income endowments of the households follow a memoryless stochastic Markov process, I argue that a risky environment where a financial catastrophe is always around the corner for any household makes these households continue to share in this income-sharing arrangement. Developing a model of informal insurance using Markov chains and running a numerical simulation, I show that the stability of such an arrangement is dependent only on the extent to which the present generation values the future alongside their degree of risk aversion and the severity of a possible catastrophe. (Sunny is now a Research Associate based in New Delhi (India) for Evidence for Policy Design in the Center for International Development at Harvard University.)

Torger Miller Computer Science major

-ADVISED BY RHYS PRICE JONES

ON THE DESIGN OF DECENTRALIZED ANONYMITY NETWORKS

Tor, the popular onion-routing based anonymity network, has an outdated threat model that does not account for the increasing surveillance capabilities prevalent today. Although certain strategies, such as traffic mixing and traffic padding, are performance prohibitive when applied at scale to the entirety of an onion-routed anonymity network, the careful application of these techniques to specific types of traffic vulnerable to low-resource attack may allow dramatically improved security with a relatively small performance cost. The Tor rendezvous model, in particular, stands to gain substantially from targeted hardening against various low-resource traffic correlation attacks.

�10


Varunavi Newar Math & Business Economics double major

-ADVISED BY JOHN RAMSAY (MATH) AND AMYAZ MOLDEDINA (BUSINESS ECON)

LEGALIZATION OF PROSTITUTION AND ITS IMPACT ON THE MARKET FOR HUMAN TRAFFICKING

Human trafficking is a form of modern slavery which involves the illegal trade of humans for the purpose of forced labour and/or sexual exploitation. Most victims of human trafficking are women and children trafficked for the purpose of sexual exploitation. Thus, the legal framework regarding prostitution may heavily influence the profitability of trafficking and the inflow of trafficked victims to a given country. This paper uses a compartmental epidemiological model to create a theoretical framework around the market for human trafficking and shows that the legalization of prostitution will lead to an increase in the volume of trafficked victims. (Varu has accepted a position as a research associate at NERA Economic Consulting, New York City)

Jacob Solomon Mathematics major (Education minor)

-ADVISED BY PAMELA PIERCE (MATH) AND MATTHEW BRODA (EDUCATION)

TO BLEND OR NOT TO BLEND: UNDERSTANDING THE IMPACT OF TECHNOLOGY IN MATHEMATICS EDUCATION

With the use of technology continually increasing in the classroom, different learning models with technology as the central component are increasing in the education world. One of these models is the blended learning model. However, the direct-teaching model continues to be widely used, even with computer usage in classrooms on the rise. This thesis was designed to find if one of these teaching models yields better academic results. Also, students’ perceptions of mathematics education were taken into account by surveys, Four Algebra I classes were split into two groups for one unit of learning. One group was taught via blended learning, while the other groups was taught via direct teaching. Students completed quizzes, a mid-unit test, and an end-of-unit exam over the duration of the study. When the findings showed that academic results did not differ between the two models significantly, more analysis was conducted using the perception surveys. The results from the perception surveys indicated that students’ self-efficacy in the mathematical ability and history of success in mathematical classes showed the greatest impact on the end-of-unit exam. This thesis also provides background in both models, as well as a variety of studies based around technology in education. Also, this thesis will describe how different educational theorists and theories, as well as state and national mathematics education standards, provide a framework to the blended learning model. Lastly, this thesis provides implications for people in education, we well as future researchers.

�11


Melissa Griffith Math and Philosophy double major

-ADVISED BY PAMELA PIERCE (MATH) AND ELIZABETH SCHILTZ (PHILOSOPHY)

THE INEFFABILITY OF “NOTHING”

Mathematics and philosophy, as disciplines, have always been intertwined through their common origin in ancient academic tradition. In these intersections, they shape one another’s conceptions of things like logic, existence, and nothingness. This Independent Study focuses upon mathematical and philosophical accounts of nothingness in order to expose the pervasiveness of ideas of nothingness throughout human intellectual tradition, as well as throughout one’s human life. Beginning with a historical account of conceptions of zero in Indian, Egyptian, Greek and Chinese mathematics, several conceptual problems with talking about and signifying nothingness are revealed. Then, the focus of the project moves to a preliminary account of the concept of emptiness in Buddhism; addresses mathematical implications of treating zero as an arithmetic number through zero divisors and division by zero; and explores Kierkegaard’s dread and despair as a fundamental, early concept in European existentialism. In conclusion, Zen Buddhist and Daoist accounts respond to all aforementioned conceptions of nothingness.

Sue Reon Kim Computer Science major


BRINGING LIFE TO A HISTORICAL DRAWING

This independent study investigates Maya’s capabilities, a computer graphics software. It discusses modeling, rigging, texturing, animation, rendering, and many other effects. It also explains how Maya’s script language (MEL) can be used to create customized computer generated graphics. In addition,it created a 3D animation based on a historic 2D Korean drawing, The Scenery of Dano Day, by Shin Yun-Bok. This involves several steps including: using Photoshop for base images; exporting these images to Maya for 3D modeling; generating a character; animating movement with Maya GUI and script language; rendering the scene; and video editing with Premiere Pro.

�12


Arpan Roy Computer Science major


LOOK ALIVE! MODELING AND ANIMATING FACIAL EXPRESSIONS OF EMOTION

Facial modeling and animation play a crucial role in today’s society in the fields of simulation, communication, and digital entertainment. Realism and efficiency are the two factors holding us back from optimizing our models of the human face. Facial expressions play a key role in human society and are strongly linked to emotion. In this paper, we explore the different methods of modeling and animating human facial expressions, highlighting the problems and solutions that have been developed. We also investigate the biological and psychological properties of facial expressions to gain a better understanding of how to induce more realism into our models. Some current research and applications in the field are explored. We apply the theory by attempting to improve on Keith Waters’ 1987 facial model to improve realism, detailing the existing model and its drawbacks as well as our improvements.

Tyler Catlin Computer Science major


PREDICTING STOCK MARKET DIRECTIONALITY USING TWITTER AND NEURAL NETWORKS

In recent years, social media has become ubiquitous in social networking and content sharing. Yet the content that is generated from these websites remains largely untapped. This Independent Study investigates the usability of Twitter data for predicting trends of share price directionality in stock markets. In particular, it uses Tweets about Tesla and Boeing together with five learning algorithms (two of them being neural networks) to predict these trends. First, the data is pre-processed and analyzed for sentiment (i.e. each Tweet is labeled as positive or negative). Then, these labeled Tweets are used together with the actual stock prices in the five learning methods. The best accuracy is obtained using a cascading neural network with a windowing technique that incorporates the share prices and the sentiment data from the five days prior to the day being predicted. The researchers found a moderate correlation (62% correct prediction) between the model’s predictions of directionality and the true directionality of the stocks sampled. Better results could be achieved with better sentiment analysis and with larger datasets.

�13


Sarah Williams Computer Science major


FROM FANTASY TO VIRTUAL REALITY: AN EXPLORATION OF MODELING, RIGGING AND ANIMATING CHARACTERS FOR VIDEO GAMES

In the last few decades, video games have quickly become one of the most popular forms of entertainment around the world. This can be linked to the improvement of computer systems and graphics which now allow for authentic and highly detailed computer generated characters. This project examines how these characters are modeled and developed. The examination of game characters entails a brief history of video games and their aesthetics. The foundations of character design are discussed and 3D modeling of a character is explored in detail. Finally, rigging or skeleton placement is investigated in order to animate the characters designed for this study. The result is two animated characters, which can be incorporated into several of the current and popular game engines. By the end to this paper, the reader should have a fundamental understanding of how a video game character is designed, modeled, rigged, and animated.

Pratistha Bhandari Computer Science major


3D MODELING AND SCENE DESIGN FOR GAMES USING UNREAL ENGINE 4

This paper outlines the process of 3D scene design through the study of lighting, shading, and texture mapping. These factors are interlinked and can merge together in complex ways to create extraordinary video game scenes. The use of various illumination models, shading models, and mapping methods are discussed. For the software portion of this project, Blender and Unreal Engine 4 are used to create a semi-realistic fantasy forest scene. Capabilities and limitations of these tools are further examined.

�14


Dagmawi Zegeye Computer Science major


AN EXPOSITION OF SUPPORT VECTOR MACHINE CLASSIFICATION WITH A PURPOSE-BUILT GUI FOR EXPLORING SOME COMMONLY USED KERNELS

We provide an exposition of the support vector machine classifier (SVMC) algorithm. We show that SVMCs produce a linear discriminant, specifically the maximal margin hyperplane and we detail how we can use the method of Lagrange multipliers and the Kuhn-Tucker conditions to find this maximal margin hyperplane. We also present some of the most popular kernels: the Gaussian radial basis function kernel, the inhomogeneous polynomial kernel, and the

hyperbolic tangent kernel and discuss their advantages in the context of SVMCs. To provide an example of SVMCs in use, we perform a series of experiments on the Titanic data set, made available by Kaggle. Finally, we present SVMC Visualizer, a web-based graphical user interface we developed in the R programming language, that can be used for experimenting with SVMCs under various feature maps and kernels on user specified datasets. We developed SVMC Visualizer with the intention that it aids users in better understanding the support vector machine classification.

Peter Nelson Computer Science major

-ADVISED BY SOFIA VISA NEEDLES AND HAYSTACK: A COMPUTATIONAL APPROACH TO FINDING SUBALGEBRAS OF THE SYMMETRIC GROUP

In abstract algebra, the symmetric group Sn on a finite set X (containing n elements) is the set of all possible mappings from X to itself, and whose group operation is function composition. In practice, Sn = all possible permutations of set X’s n distinct elements; thus the magnitude of Sn is n!. We can partition Sn into one or more parts (also known as blocks or cells); that is to say, we can divide Sn into a union of overlapping, non-empty subsets; these blocks form a basis for Sn. We can also say that they form a subalgebra of Sn if any combination of these parts via multiplication produces a linear combination of the blocks formed by the initial partitioning. This paper seeks to find out, for any given n, how many of the potential Bell number of ways to partition Sn actually yield a working subalgebra. Once all of the subalgebras for n = 1,2,… are found, the intention is to try to discern if there are any mathematical patterns and/or categories that these partitions fall into. The number of possible ways to partition Sn increases dramatically as n increases, meaning that even for n = 4 we are already looking at a 15-digit number of possible ways to divide Sn. This means that checking all possible partitions of Sn via brute force would be prohibitively inefficient. Instead, this paper will attempt to find all possible subalgebras via several different approaches, including Monte Carlo simulations, as well as examining subsets of prohibitively large search spaces.

�15


William Rial Computer Science & Classical Studies major

-ADVISED BY SOFIA VISA (CS) AND MONICA FLORENCE (CLASSICAL STUDIES)

VENI VIDI DIDICI: A VIDEO GAME FOR LEARNING LATIN

Learning a language is a difficult task and often requires many years of dedicated practice to master. Thus, many have attempted to find other ways to learn a language. One possible solution is through games. J.P. Gee writes that playing a video game

actually teaches a variety of skills based on the challenges presented. Websites like Dueling reward the user for practicing a language and offer greater prizes for regular participation. Some other attempts emphasize fun but neglect the pedagogical aspect. This project attempts to create a video game that is both educational and engaging. The game, called Veni Vidi Didici, has two distinct parts: a translation game inspired by Dueling, and an interactive game where the user plays as a Roman legionary in an episode

from Caesar’s De Bello Gallico. In part one, we explain why choosing Caesar and his narrative about the Gallic war benefits eager students. Then we examine Caesar’s life up to the Gallic War and his habits as a general before we consider his historiography and its implications. In part two, we discuss the Unity Game Engine and the process for creating the game.

Collaborating Digitally: Engaging Students in Public Scholarship, Bucknell University, 2015

William Rial presented his interdisciplinary summer research “A Framework for Using Tesseract to Transcribing Early Modern Texts Having Non-standard Fonts”. He gave an oral and a poster presentation on how he used machine learning and optical character recognition techniques to automate the transcription of two French volumes from the 17th century (Conversations sur différents sujets, by

Madeleine de Scudéry, 1680).

�16

Sample of original text as image fileResulted transcription in a text file


MATH/CS FACULTY

�17

Dr. Hartman at the Senior I.S. pizza party with his advisee, Sunny Mitra ’16.

Dr. Bowen met up with Math/CS alums Evan Radkoff ’12 and Itai Njanji ’11 in Seattle at the Joint Meeting of the AMS/MAA

Jennifer Bowen Associate Professor of Mathematics & Chairperson

PH.D. VIRGINIA 2005

Teaching 2015-2016: Abstract Algebra, Mulltivariate Calculus, First Year Seminar, 2 Senior Independent Study advisees

Dr. Bowen was accepted into the Research Experience for Undergraduate Faculty on the Mathematics of Data at Duke University for July 2016 (as a part of the Information Initiative at Duke and the American Institute of Mathematics). She remains engaged in the greater mathematical community, serving on the Mathematical Association of America’s Problem Books Series Editorial Board and continuing to co-organize and co-facilitate the Graduate Education Mentoring Workshop day for the Summer Mathematics Program for Women alumnae each January at the national Joint Mathematics Meetings. Locally, she advised multiple AMRE projects this past summer and serves on Wooster’s STEM Advisory Board, consulting on STEM Zone and introductory STEM course pedagogy at the College. Her First Year Seminar course, “The Signal and the Noise: Why Numeracy Really Matters”, focused on quantitative literacy.

James Hartman Professor of Mathematics

PH.D. MICHIGAN STATE 1981

Teaching 2015-2016: Linear Algebra, Advanced Linear Algebra, Probability and Statistics I, Calculus with Algebra A, 2 Senior Independent Study advisees

Dr. Hartman attended the National Joint AMS/MAA meeting in Seattle in January, where he presented “Learning to Ask Questions: A Matrix Project”. He also presented “Day with a Reader” at the Advanced Placement Annual Conference last fall. He served as an AP Calculus exam leader and AP Summer Institute instructor.


MATH/CS FACULTY John Ramsay Professor of Mathematics and Associate Dean for Experiential Learning

PH.D. WISCONSIN 1987

Teaching 2015-2016: Operations Research, Topology, 1 Senior Independent Study advisee

In addition to his math teaching and experiential learning duties, Dr. Ramsay served as director of the 2016 AMRE program which employed a record-setting 41 students. He was very honored to receive the 2016 award for Outstanding Leader in Experiential Education in Higher Education from the National Society for Experiential Education at the 44th Annual NSEE Conference in St. Pete Beach, Florida. Dr. Ramsay’s father was one of the founders of the NSEE back in 1971.

Pamela Pierce Professor of Mathematics

PH.D. SYRACUSE 1994

Teaching 2015-2016: Calculus I, Problem Seminar, Real Analysis, Transition to Advanced Mathematics, 3 Senior Independent Study advisees

Dr. Pierce traveled to Europe this summer and attended the Summer Symposium in Real Analysis in Sarajevo. She then traveled to Prague where she worked on a research project with Dr. Ondrej Zindulka (Czech Technical University) and Dr. Bruce Hanson (St. Olaf College). Dr. Pierce is excited that Dr. Zindulka will be visiting Wooster next year, because this will provide them with an unusual opportunity to continue research throughout the academic year. She gave colloquium talks at Hamilton College and Oberlin College on “Circle Squaring and Undergraduate Research”. She attended the Joint Meeting of the AMS/MAA and presented “Learning to Ask Questions: A Matrix Project” along with Jim Hartman.

�18

Dr. Ramsay and some knot theory student researchers


MATH/CS FACULTY

�19

Drew Pasteur Associate Professor of Mathematics

PH.D. NORTH CAROLINA STATE 2008

On Sabbatical 2015-2016

Dr. Pasteur spent his research leave primarily doing sports analytics research, specifically in developing improved measures for evaluating NFL punters and kickers. On a fall visit to collaborator (and former department member) Dr. John David, he gave talks to students at both Virginia Military Institute and Roanoke College. Dr. Pasteur and Dr. David organized the "Mathematics and Sports" session at the Joint Meetings of the AMS/MAA, and they will have a chapter on "Evaluation of Quarterbacks and Kickers" in the forthcoming CRC Press book, "Handbook of Statistical Methods and Analyses in Sports”. This past summer, Dr. Pasteur visited Ashesi University in Ghana, together with two other Wooster faculty members, to lead a three-day faculty development workshop, which included significant discussion of AMRE and I.S.

Robert Wooster Assistant Professor of Mathematics

PH.D. CONNECTICUT 2009

Teaching 2015-2016: First Year Seminar, Numerical Analysis, Calculus II, Probability and Statistics II, 3 Senior Independent Study advisees

In addition to advising two of the math modeling contest teams and supervising two AMRE projects, Dr. Wooster served on the Strategic Planning and Priorities Advisory Committee. His First Year Seminar, “Freedom in the Digital Age”, focused on recently seen challenges to freedom and liberty, from issues of online privacy to “PC” restricted speech on many college campuses. The class read works by George Orwell, Greg Lukianoff, Thomas Sowell, and others.

Sadly, Dr. Wooster and his wife Krista have decided to leave Wooster and return to their home state of Connecticut. We wish him all the best and will miss the practical jokes that he played on his department colleagues (as well as the ones that were played on him).


MATH/CS FACULTY

�20

Denise Byrnes Associate Professor of Computer Science

PH.D. OHIO STATE 1992

Teaching 2015-21016: Multimedia Computing, Theory of Computation, Computer Graphics, Algorithm Analysis, Data Structures and Algorithms, tutorial in Designing Virtual Reality Environments, 5 Senior Independent Study advisees

Dr. Byrnes served on the program committee and was a paper reviewer for the Midstates Conference on Undergraduate Research in Computer Science and Mathematics and also the Ohio Celebration of Women in Computing. She participated in a summer meeting of the Liberal Arts Computer Science Consortium Working Group at Williams College.

Sofia Visa Associate Professor of Computer Science

PH.D. CINCINNATI 2007

Teaching 2015-2016: Scientific Computing, Data Structures and Algorithms, Computer Networks and Communication, 7 Senior Independent Study advisees

Dr. Visa and CS major William Rial ’16 worked on an interdisciplinary research project together with Laura Burch from the French Department and Jacob Heil, a Mellon Digital Scholar at Wooster, on creating an optical character recognition software that automatically transcribes French text from pdf files to text files. This work was presented at digital collaboration conference at Bucknell University and at the MCURCSM conference at Bowling Green. On campus, Dr. Visa participated in the Expanding Your Horizons workshop for middle school girls, teaching three programming sessions in Python.

Rhys Price Jones Pocock Family Distinguished Visiting Professor of Computer Science

PH.D. LONDON 1976

Teaching 2015-2016: Imperative Problem Solving, Data Structures lab, Principles of Computer Organization, Software Engineering - Databases, 3 Senior Independent Study advisees

Dr. Price Jones came to Wooster on a one-year appointment after several years of teaching at George Washington University and Oberlin, among others. He is the co-author of a textbook entitled BioInformatics: A Computing Perspective which was published in 2008 by McGraw-Hill.


MATH/CS FACULTY

�21

Matthew Moynihan Visiting Assistant Professor of Mathematics

PH.D. BRANDEIS 2012

Teaching 2015-21016: Calculus I and II, Differential Equations, Combinatorics and Graph Theory, Problem Seminar, 3 Senior Independent Study advisees

Dr. Moynihan served as the Math Club advisor and on the College’s Judicial Board. He advised two of the math modeling contest teams and two AMRE projects (Goodyear and Artiflex Manufacturing). He worked on a research project investigating Gellel’s Conjecture concerning descents and inverse descents of permutations.

Mary Jo Kreuzman Assistant Professor of Mathematics

PH.D. NOTRE DAME 1985

Teaching 2015-2016: Math in Contemporary Society, Calculus I, Transition to Advanced Mathematics

Ronda Kirsch Math Center Coordinator and Instructor of Mathematics

M.A. KENT STATE UNIVERSITY 2010

Teaching 2015-2016: Calculus with Algebra A

Ms. Kirsch attended the National Council of Teachers of Mathematics Annual Meeting in San Francisco. The NCTM s the leading professional organization for teachers of mathematics and provides broad national leadership in matters related to mathematics education. Ms. Kirsch taught the math component of the College’s Youngstown Early Intervention Program and advised one AMRE project for Progressive Insurance.


MATH/CS COLLOQUIUM SERIES September 3, 2015 S-Stem Scholar/Summer Research Experiences Part 1 (Michael Bush, math/physics ‘16, Matt King-Smith, physics ‘16, William Rial, CS/classical studies ‘16)

September 10, 2015 S-Stem Scholar/Summer Research Experiences Part 2 (Nate Johnson, physics ’16, Mel Griffith, math ’16, Lindsay Robinson, chemistry ’17)

October 22, 2015 Archimedes meets geometric series (Dr. Robert Kantrowitz, Professor of Mathematics, Hamilton College)

October 27, 2015 Modeling Insect Olfaction: How Bugs Smell (Dr. Pamela Pyzza, Department of Mathematics, Ohio Wesleyan University)

November 12, 2015 The Nova Platform: Why Programming is More than Code (Dr. Richard Salter, Professor of Computer Science, Oberlin College)

February 25, 2015 Along Came Polly, or, what do you want to do with a math degree? (Ryan Snyder, Actuary, Wooster Math Major, Class of 2014)

�22

2015-16Math/Computer Science

SeriesThe Nova Platform:

Why Programming is More than Code

Dr. Richard SalterProfessor of Computer Science

Oberlin College

Thursday November 12

11:00 am

Taylor 110

Pizza lunch to follow

Nova is a new Java-based modeling platform that naturally supports the creation of models in thesystem dynamics, spatial and agent-based modeling paradigms. Nova uses a visual language to express model design, and provides automatic conversion for such models to script form for execution. Nova’s runtime operates either as desktop application, or using a Web-based runtime engine on either the browser or server side. Nova’s architecture promotes hierarchical design, code reuse, and extensibility through the use of plug-ins. In this talk I will demonstrate the Nova system with a particular emphasis on how its design impacts its expressive power.


HONORS & AWARDS TO MATH/CS MAJORS

�23

Latin Honors Summa cum laude Kiera Dobbs

Magna cum laudeMichael BushAdam CoppockKatelyn FrenchSayantan MitraParoma PalchoudhuriWilliam Rial

Cum laude Pratistha BhandariMelissa GriffithSue Reon KimTorger MillerVarunavi NewarArpan RoySarah Williams

Phi Beta Kappa Kiera DobbsParoma Palchoudhuri

The Dan F. Lockhart Outstanding Senior Award Sayantan Mitra

The William H. Wilson Prize in Mathematics Michael Bush

The Foster Prize in Mathematics Adam Coppock

J. Howard Morris and Josephine Morris Volunteer Service Award

Michael Bush

Elizabeth Sidwell Prize in Mathematics Kiera Dobbs

The Lyman C. Knight Sr. Prize in Mathematics and Physical Education

William Koenig

The William H. Wilson Prize in Mathematics Michael Bush

International Paper Company Foundation Business Economics Prize

Varunavi Newar

The Edward Taylor Prize Emily Howerton

D. Ivan Dykstra Philosophy Roundtable Book Prize

Melissa Griffith

Elias Compton First Year Prize Avi Vajpeyi

Tom Newiswander Memorial Award Sayantan Mitra

Adam Coppock, Kiera Dobbs, and William Koenig at the Recognition Celebration held in February.


CONTESTS & CONFERENCES

�24

Senior CS/Classics major William Rial attended the 2015 MCURCSM conference at Bowling Green State University and presented “A Framework for Using Tesseract to Transcribing Early Modern Texts Having Non-standard Fonts” along with Professor Sofia Visa.

The 2015 ACM ICPC East Central North America Regional Programming Contest took place in October at Youngstown State University.

Wooster’s programming team “StraightOttaWooster”, consisting of Nan Jiang, Dagmawi Zegeye, and Xiangyu Li, and advised by Professor Visa, placed 58th out of 130 participating teams. Considering only institutions similar to Wooster and excluding large universities, our team placed 14th.

National Meeting of the AMS/MAA Wooster students Kiera Dobbs, Brian Foley, and Shiwani Varal joined faculty members Drew Pasteur, Pam Pierce, John Ramsay, Jim Hartman, and Jen Bowen in Seattle for the joint AMS/MAA conference. Kiera Dobbs, a senior, presented “Optimal Placement of Family Planning Centers,” which highlighted some of the results of her Senior Independent Study. Brian Foley gave a talk titled “Solving the Unknotting Problem,” based on a research project that he worked on during the summer AMRE (Applied Methods and Research Experience) program.


AMRE

�25

Western Reserve Group Data Mining Automation for Auto Insurance Data

Team members: Scarlett Chen, Khoa Nguyen, Ralph Xu

Faculty advisors: Dr. Jennifer Bowen & Dr. Robert Wooster (Math)

Project Description: The goal of this AMRE project for Western Reserve Group is to automate the process of data extraction from Microsoft Access into Excel and to structure the data appropriately for future analysis in tracking losses and setting premium rates. We are given several Access tables and two Excel files that contain the necessary data and calculations for auto insurance. By creating new queries to establish the data connections between Access and Excel, and configuring the calculations to adapt to the dynamic data, users can now interact with the data in the Excel files directly, hence increasing the ease and efficiency of their work.

Applied Methods & Research Experience

Artiflex Manufacturing Improving Revenue Forecasting and Creating a Capacity Forecasting Model Team members: James Bai, Efua Bainson, Aidan Brown, Conor Maley

Faculty advisors: Dr. Drew Pasteur & Dr. Matthew Moynihan (Math)

Project Description: ArtiFlex Manufacturing produces replacement automobile parts. The AMRE team was hired by ArtiFlex to create a capacity forecasting model as well as improve the existing revenue forecasting model. These models are designed to help ArtiFlex make predictions about future business opportunities and expansion needs. Using Excel, the team refined the user interface and created additional data management features in the revenue forecasting model. The capacity forecasting model was designed to have a user friendly layout and a color coded warning system for machines that are near or over capacity. To estimate how much capacity has been used, the model automatically calculates the hours needed to complete different contracts, and compares these values to the total possible hours per month. Predictions become more accurate as more data becomes available on a certain contract.


AMRE

�26


The College of Wooster - Experiential Learning Student activity data for experiential and student affairs assessment Team members: JoJo Tang, Jena Styka, Lissette Torres

Faculty advisors: Dr. John Ramsay (EL) and Missy Schen (Assessment)

Project Description: The Experiential Learning team was tasked with three deliverables to help The College of Wooster rethink the way experiential learning (EL) is talked about and assessed on campus. First, we created a common definition of EL, which we hope will help unite the College in thinking about EL as a campus-wide goal. Second, we developed a framework for the College to use to talk about an EL experience and to assess the value it has on student development. Lastly, we investigated how the College could track student involvement in co-curriculars. These deliverables can help the College improve EL on campus, as well as help with student development, marketing, retention, and alumni engagement.

The College of Wooster - Predicting Retention Systematic Approach to Assessing Student Retention Probability Team members: Jack Marousek, Abbey Partika, Unnati Singhania

Faculty advisor: Dr. Brian Karazsia (Academic Advising & Psychology)

Project Description: The AMRE Retention team worked to identify variables that could improve retention prediction at The College of Wooster. A mixed method approach was used which included conducting a literature search and having departmental meetings. They recommended that the College switch to a mechanical or systematic method of predicting retention instead of a clinical approach attributed, which is prone to human error. Non-academic predictor variables were the focus of the report which was backed by empirical data. The report also contained a series of measurements for the variables. These measurements can be implemented in the future to collect accurate data and generate a systematic equation in to predict the likelihood of a student leaving the College before graduation.


AMRE

�27

Applied Methods & Research ExperienceThe College of Wooster Alumni Board Storycorp 150 Voices Team members: Sam Carlson, Sarah Torio, Ronnie Wright

Faculty advisor: Dr. Richard Figge, Emeritus

Project Description: For this project, we produced, wrote, and edited over 50 interviews for podcasts lasting 3-12 minutes long. To produce these casts, we contacted hundreds of alumni and faculty and staff. In interviews we asked a general set of questions while having specific questions designed for each one of our interviewees. We edited out the “fat” and kept the “meatier” parts of the interviews. The results of these methods gave us 76 interviewees, and over 90 podcasts with thought provoking answers and heart felt responses. The original name of the project, The College of Wooster Alumni Board Storycorps AMRE project, was changed to Project 150 (Pronounced: Project one-fifty).

The College of Wooster - Advising Tool Advising Project in Design Thinking Team members: Aaron Brown, Paige McKean, Thanh Nguyen

Faculty advisors: Michelle Johnson (Communication) and Alison Schmidt (Education)

Project Description: The goal for the team was to use a design thinking model to provide information regarding how to improve the advising process at The College of Wooster so that students cultivate the Graduate Qualities. Following the design thinking model, the team conducted interviews with faculty, staff, and students to begin to identify common themes with the advising process. Themes from these interviews served as the basis for a survey that was sent to the entire College of Wooster student body. After reviewing data from the survey in conjunction with interview responses and personal reflection, the team identified the main challenges with the current advising process at Wooster and provided recommendations to overcome those challenges. Along with a final report detailing their process and conclusions, the team also compiled an annotated bibliography of sources relating to the design thinking process. The team’s work will be the basis for the future development of a new online advising tool.


AMRE

�28

Digital History - Tizzano Museum Digital Archiving for Online Museum Team members: Anna Claspy, Theresa Dunn, Heather Smith

Faculty advisors: Katie Holt and Madonna Hettinger, History

Project Description: Dr. Anthony Tizzano, an Ob/Gyn for the Cleveland Clinic, owns a collection of over 2400 gynecological and obstetrical antiques. He hopes to bring this collection to a wider audience and provide a resource for the history of women’s health. Dr. Tizzano hired a research team with digital history expertise to help organize and curate a digital museum to display his collection. For a nominal fee, site visitors can download high-resolution PowerPoint slides for lectures or personal use. Dr. Tizzano provided us with PowerPoint files from which we created individual PowerPoint presentation files for each object. SquareSpace provided a user-friendly platform on which we could easily create click-through pages and add images. On SquareSpace, we created section and subsection pages for the collections. We created product pages to display individual objects and allow patrons to click for further information. Interested patrons can choose to “Add to Cart,” which allows them to purchase a PowerPoint presentation that includes more information on the object, access to the image, and references. Drawing on our experience with multimedia composition, we took on an additional project of creating e-books that patrons can purchase for a higher price, but with more content and information.

The College of Wooster - Entrepreneurship Entrepreneurial Ecosystem Analysis Team members: Spencer Gilbert, Kayleigh Marks, Emily Partika

Faculty advisors: Peter Abramo, Entrepreneurship and Brett Woodard, Experiential Learning

Project Description: The College of Wooster’s Center for Entrepreneurship initiated a project in 2016 to develop a plan to build a sustainable college-community entrepreneurial ecosystem. Based on preliminary research conducted, the team accomplished the project objective by developing a plan to create an active, interconnected local ecosystem that supports entrepreneurship and ties The College into various elements provided by community institutions. After researching, conducting interviews with local entrepreneurs, and attending the Deshpande Symposium at UMass Lowell, the team developed project ideas to implement at The College of Wooster and proposed them at the end of the eight-week period. The following are a sample of their ideas: a new group they referred to as “IDEA” (Innovation and Development of the Entrepreneurial Arts); an entrepreneurial living learning community; and a consulting course that would represent a lesser capacity of the AMRE consulting program, and would be tied to a class for credit.



AMRE

�29


Goodyear Tire & Rubber Company - OTR OTR Mine Site Severity Index Team members: Jacob Denbeaux, Colby Jeffries, Robin Morillo, Rahab Wangari

Faculty advisors: Drew Pasteur and Robert Wooster (Math)

Project Description: The goal of this project was to build an off the road (OTR) tire performance prediction model. Because OTR tires are expensive, Goodyear wanted to know how well tires will perform at a particular mine site before selling them to the mine site owners. The AMRE team received 6 data sets containing GPS data from various mine sites as well as supplemental information about the mine sites. The final result of this project was a mine site analysis tool with a built-in prediction model created by the team. The mine site analysis tool helps the company compare tire performance in different mines, predict tire performance in future mine sites, and visualize the GPS data in new and useful ways.

Goodyear Tire & Rubber Company - Retread Manufacturing Footprint vs. Market Share Optimization Team members: Makafui Fie, Brian Foley, Nan Jiang, Yunjia Zeng

Faculty advisors: Jennifer Bowen and Matt Moynihan (Math)

Project Description: The goal of our project is to summarize, classify, and visualize all the data files that The Goodyear Tire and Rubber Company provided in May 2016, in order to help them with their retread business. We were given several Microsoft Excel files containing data about Goodyear owned and Goodyear authorized retreading plants. We were also given some competitor data, and a sample of truck fleet data in Florida.Our project has two main deliverables:1. A web application that reads data from files and projects the data

onto a map. The map must be interactive, easy to use, and show data in such a way that conclusions can be easily drawn from the map.

2. A database in Microsoft Access that contains all of the original data provided, which is organized and easy to maintain, and fits the structure of Goodyear’s data.


AMRE

�30

Progressive Insurance Competitor Analysis Team members: Gina Lam, Ayeley Commodore Mensah, Tyson Vogel, Claire Vu

Faculty advisors: John Ramsay and Ronda Kirsch (Math)

Project Description: Progressive Insurance’s Competitive Intelligence (CI) Team was interested in exploring how competitive their 2015 auto insurance quotes were against the following competition: GMAC, Sentry, Infinity, and Zurich. The College of Wooster AMRE Team was provided with four main competitor .csv files along with 6 additional .csv files which included: Coverage Response Table 1, Coverage Response Table 2, Driver Table, Vehicle Table, Violation Table and Quomation Violation Table. In addition to those ten files, the team was provided with a glossary that listed all of the variables contained in each file along with instructions for merging each file. Various forms of analysis were used to evaluate Progressive’s competitiveness:* Logistic regression* Multivariate head to head analysis* Variation of winning percentage by area or over time* Performance vs. multiple carriers* Quote mixEach of these methods offered different insight into Progressive’s competitiveness and the variables that had an effect on it.


MATH CLUB (advised by Prof. Matt Moynihan) Leadership Team 2015-2016Michael BushSunny MitraShiwani Varal

Math Club sponsored several game nights and talks and brought back the long-absent “Problem of the Week” which had great participation. Other events included the annual Math/CS Fall Picnic and Taylor Bowl 27.

Math/CS 103.8 vs. Physics 96.3

Math/CS remains Taylor Bowl champions!

2015-2016 department newsletter

Documents