[Amath-seminars] *Tomorrow* Computational Finance Seminar on May 4 (Fri), 11--11:50am

Laurie Feldman lf23 at uw.edu
Thu May 3 12:39:42 PDT 2018

Hi All,

We have a special Computational Finance seminar on May 4 (Fri), 11--11:50am
in Wan Conference Room inside Lewis Hall. Our speaker is Patricia Ning from
the Dept. of Statistics and Applied Probability at UCSB, and she'll tell us
about Multivariate Bayesian Structural Time Series Model and its
Applications on Finance. See below for abstract. Hope to see some of you

-Tim Leung

*Speaker: *Patricia Ning, Dept. of Statistics and Applied Probability, UCSB


*Title:* Multivariate Bayesian Structural Time Series Model and its
Applications on Finance

*Abstract:* Bayesian structural time series (BSTS) model is a machine
learning technique used for feature selection, time series forecasting,
nowcasting, inferring causal impact and other, which was developed
by Steven Scott (Senior Economic Analyst at Google) and Hal Varian (Chief
Economist at Google). We extended it to its complete multivariate
version: Multivariate Bayesian Structural Time Series (MBSTS) Model. An
empirical study with one-step-ahead prediction on the max log return of a
portfolio of stocks that involves Bank of America (BOA), Capital One
Financial Corporation (COF), J.P. Morgan (JPM) and Wells Fargo
(WFC), confirmed that this multivariate model outperforms three other
benchmark models, viz. a model that treats each target series as
independent, the autoregressive integrated moving average model with
regression (ARIMAX), and the multivariate ARIMAX (MARIMAX) model. We also
propose a modified MBSTS model with text data mining, which is fully tested
on the stock price time series of two competing electronic commerce
companies Amazon and Ebay, with Sentiment Analysis on financial news and
twitter feeds.

(This talk is based on joint work with S. Rao Jammalamadaka and Jinwen Qiu)


Program Manager / Computational Finance & Risk Management
Applied Mathematics

Lewis Hall Box 353925
Seattle, WA 98195
lf23 at uw.edu / washington.edu

[image: http://www.washington.edu/brand/files/2014/10/e-sig.gif]
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman13.u.washington.edu/pipermail/amath-seminars/attachments/20180503/bda2c8fe/attachment.html>

More information about the Amath-seminars mailing list