Research Experiences For Undergraduates (REU)

REU 2007 Projects

Modeling Stock Returns and Optimizing Bond Portfolios

Sponsor: State Street Global Advisors (SSgA)
Advisor: Prof. Bogdan Doytchinov


Naomi Brownstein		Patrick Crutcher		Yu-Jay Huo		Sean Skwerer		Grant Weller

State Street Global Advisors (SSgA) is the largest institutional asset manager in the world. As such, SSgA is interested in modeling monthly stock returns and managing optimal bond portfolios. We will all be working on two related projects:

Stocks project: Develop a model to evaluate and rank future performance of stocks, based on some known factors, such as value, growth, sentiment, and quality. Since the data is extremely noisy and will likely cause the model to change from month to month, we are also interested is making this model smooth over time to minimize stock trading costs. Regression will be our primary tool for model building; however, alternatives such as artificial neural networks will be examined as well.
Bonds project: A bond is a measure of debt issued by an institution. Investors commonly include bonds in their portfolios due to the fact that they are less volatile than stocks. However, bonds are associated with a risk of default, measured by credit ranking agencies, from AAA -- least risky to D -- most risky. This project will analyze investment grade bonds: A and BBB bonds issued by non-financial companies. For each bond issuer, State Street has computed a score, which tries to predict how well the bond will perform compared to treasury bonds. Given 160 months of benchmark portfolio weights, returns, and these scores, we seek to construct a portfolio that beats the corporate bond index. To verify the results, comparisons of our rankings against the benchmark returns and simulations will be conducted.

Mathematical Model of the Self-Tapping Screw Insertion Process

Sponsor: Bose Corporation
Advisor: Prof. Suzanne Weekes


Jonathan Adler		Matthew Bader

A self-tapping screw is a high-strength one-piece fastener that is driven into preformed holes. The goal of the Major Qualifying Project (MQP) completed by A. Leo, et al. and the Research Experience for Undergraduates (REU2006) completed by Miller, et al. was to create a mathematical model that allows users to input data about their self-tapping screw and the material it is entering and output a torque curve which models the fastening process. We improve the algorithm to include the modeling of the failure of the joint as well as a model for the clamp load of the joint. We will also develop a model of heat dynamics in the screw that considers the rate of screw insertion, i.e. RPM.

Quantifying Uncertainty in Predictions of Hepatic Clearance

Sponsor: Pfizer
Advisor: Prof. Jayson D. Wilbur


Paul Bernhardt		Jaye Bupp		Morgan Gieseke		Nathan Langholz		Christopher Steiner

In the process of drug development, it is important to understand the pharmacokinetics of candidate compounds. Among the pharmacokinetic parameters of interest is hepatic clearance, which is the rate at which blood is cleared of the drug by the liver. Generally, point estimates for hepatic clearance are obtained from preclinical data and used in clinical models without accounting for the uncertainty in these estimates. The goal of this project is to construct a Bayesian model for hepatic clearance which can be used to quantify the uncertainty in terms of a posterior distribution and evaluate the sensitivity of the distribution to various model assumptions.

Maintained by webmaster@wpi.edu
Last modified: Jun 21, 2010, 03:04 UTC