Research Experiences For Undergraduates (REU)

REU 2015 Projects

Predictive Features of Carbon Nanotubes from Factors of Production

Industrial Sponsor: Dr. Mark Banash - Nanocomp Technologies, Inc.

Faculty Advisor: Prof. Matthew Willyard


Iris Bennett		Zaynab Diallo		Tiffany Sunderland		Matthew Quinn

Carbon nanotubes are cylindrical structures made from carbon atoms that are only billionths of a meter in diameter. Though they have only been widely studied since 1991, carbon nanotubes have already been incorporated into a variety of industries for the improved strength, conductivity, and resilience they offer over other materials. At a submicroscopic level, these nanotubes can be classified into three distinct types which are differentiated by their amounts of curvature and bundling. However, it is not known what combination of factors of production results in a given type, as the mechanism behind their manufacturing is not yet fully understood. Our goal is to establish relationships between factors and their resulting varieties of nanotubes. Using scanning electron microscope and transmission electron microscope images provided by our sponsor, Nanocomp Technologies, Inc., we will use supervised learning to automate the classification of the images into the three categories. Then, we will use classification models such as decision trees, clustering, and logistic regression to predict the variety of nanotubes that will result from a given set of inputs.

Integrating Fish Movement in Multispecies Population Models.

Industrial Sponsor: Dr. Sarah Gaichas, NOAA Northeast Fisheries Science Center

Faculty Advisors: Prof. Minghao Wu Rostami and Prof. Burt Tilley


Cody FitzGerald		Alexis Gambino		James Keane		China Mauck

Fisheries management in the US is currently based primarily on spatially aggregated single-species analyses. The goal of these analyses is to estimate population status relative to management objectives for sustainable stock yield. If sufficient data are available, single species population dynamics models are used to estimate current stock status and expected productivity under varying levels of fishing mortality. These models are often quite complex in terms of accounting for population age structure and employing statistical parameter estimation with multiple data sources. However, they still cannot evaluate population productivity changes due to multispecies interactions, changes in migratory habits, and or changes in species distributions. In particular, climate change can drive changes in species distributions, resulting in different patterns of species overlap and potential changes in population productivity and species interactions. In this project, we would like to explore multiple ways to estimate fish movement from various data sources, as well as how best to incorporate this information into multispecies population models in this region.

ETF-Based Models for Liquidity Risk

Industrial Sponsor: Dr. Ritripua Samanta - State Street Global Advisors

Faculty Advisor: Prof. Marcel Blais


Jonathan Beall		Kelli Dowd		Rivers Jenkins		Zoe Rehnberg

An exchange traded fund (ETF) is a security that represents an underlying collection of assets and trades with stocks throughout the day on an exchange. ETFs provide easy diversification for investors as well as a more liquid market for positions dependent on the underlying assets. Liquidity in these ETF markets is an important issue for financial institutions like State Street Global Advisors (SSgA) who must manage risk and participate in the buying and selling of assets in these markets. In illiquid markets, large transactions can result in high trading costs due to illiquidity or even an inability to completely liquidate the necessary assets in the necessary time frame. In order to accurately assess the liquidity in the underlying basket, these financial institutions rely on liquidity metrics that model and track changes in liquidity over time. One such metric was developed in 2014 by the Worcester Polytechnic Institute Financial Math REU team that considers differences between the ETF share price and the net asset value (NAV) or price per share of the basket of underlying assets. Because these two values are both based on the same securities, the price difference between them can be attributed solely to a difference in liquidity. The ETF-NAV Liquidity Metric (ELM) that was developed exploits this price difference to quantify the liquidity of the underlying assets. In our work, we expand ELM beyond simple considerations of ETF price and NAV to include the effects of market depth and time-to-liquidation. While expanding the model we focus on high risk, high yield bond ETFs like SSgA's JNK, though we expect that the methods can be applied to a wider range of ETFs.

Maintained by webmaster@wpi.edu
Last modified: Sep 16, 2015, 21:02 UTC