My name is Weston Barger. I’m a PhD student in Applied Mathematics at the University of Washington. My academic interests center around mathematical finance, pricing theory and asymptotics.


Optimal liquidation under stochastic price impact

We assume a continuous-time price impact model similar to Almgren-Chriss but with the added assumption that the price impact parameters are stochastic processes modeled as correlated scalar Markov diffusions. In this setting, we develop trading strategies for a trader who desires to liquidate his inventory but faces price impact as a result of his trading. For a fixed trading horizon, we perform coefficient expansion on the Hamilton-Jacobi-Bellman equation associated with the trader’s value function. The coefficient expansion yields a sequence of partial differential equations that we solve to give closed-form approximations to the value function and optimal liquidation strategy. We examine some special cases of the optimal liquidation problem and give financial interpretations of the approximate liquidation strategies in these cases. Finally, we provide numerical examples to demonstrate the effectiveness of the approximations.

The paper can be found on arXiv

Approximate pricing of European and Barrier claims in a local-stochastic volatility setting

We derive asymptotic expansions for the prices of a variety of European and barrier-style claims in a general local-stochastic volatility setting. Our method combines Taylor series expansions of the diffusion coefficients with an expansion in the correlation parameter between the underlying asset and volatility process. Rigorous accuracy results are provided for European-style claims. For barrier-style claims, we include several numerical examples to illustrate the accuracy and versatility of our approximations.

The paper can be found on arXiv.

Recommender Systems for Movie Rating Data

This machine learning project was concerned with matrix factorization techniques for recommender systems. This was completed as a final project for the UW course Machine Learning, CSE 546. The project poster can be found here (pdf), and the paper can be found here (pdf).

Predicting interest in rental listings

This machine learning project is concerned with building statistical models for predicting user interest levels in rental listing data. This was completed as a final project for the UW course Machine Learning for Big Data, CSE 547.

The code, poster, and paper can be found on github.

LaTeX style packages

Paper style package

Inspired by and developed in conjunction with Matt Lorig, this package provides concrete modern font, small cap section headers, hyperref equation numbering and a variety of other stylistic choices.

The package can be found on github.

Homework style package

This package extends the paper style package to include convenient environments for typesetting homework assignments. It also optionally modifies equation numbering to correspond to exercise number.

The package can be found on github.