Hi everyone!
This week at the Applied Statistics Workshop we will be welcoming* Dr.
Francesca Dominici*, a Professor of Biostatistics at the Harvard School of
Public Health. She will be presenting work entitled *The Health Impacts of
Climate Change: The Evidence & The Uncertainty*.
As usual, the workshop will meet in CGIS Knafel Room 354 from 12-1:30 pm.
Lunch will be provided. See everyone there!
-- Dana Higgins
Hi everyone!
This week at the Applied Statistics Workshop we will be welcoming *Yaron
Singer*, a Professor of Computer Science here at Harvard. He will be
presenting work entitled* Ceci n'est pas un #ad: Detection and Analysis of
Covert Ads in Social Media*. Please find the abstract below and on the
website (here
<http://projects.iq.harvard.edu/applied.stats.workshop-gov3009/presentations…>
).
As usual, we will meet in CGIS Knafel Room 354 and lunch will be provided.
See you all there!
-- Dana Higgins
*Title:* Ceci n'est pas un #ad: Detection and Analysis of Covert Ads in
Social Media
*Abstract:* In online social networks such as Facebook or Twitter users can
advertise products and services by using appropriate labels to convey the
content they publish is sponsored. A growing concern is that many users
intentionally conceal the commercial intent behind the promotions they
publish, disguising them as authentic recommendations.
In this talk we will discuss the problem of covert ads in social media. We
will present a systematic approach for covert ad detection in social media
platforms and use this approach to analyze covert ads. The main challenge
is due to the fact that covert ads are not labeled and it is a priori
difficult to distinguish their content from authentic recommendations. We
develop specialized algorithms for characterizing redirection links used
for sponsored referrals, as well as classifiers for detecting content that
is of commercial nature. We then apply these algorithms to detect covert
ads and study this phenomenon. To evaluate the performance of the
algorithms we use the Mechanical Turk platform which allows us to evaluate
our algorithms using human annotators.
Hi everyone!
Remember tomorrow, Wednesday November 11th, we are welcoming Devin Caughey
to present Bayesian Population Interpolation and Lasso-Based Target
Selection in Survey Weighting. He wanted to circulate the paper before the
presentation in case anyone was interested. Please find it attached.
See you all tomorrow!
--- Dana Higgins
Hi everyone!
This week in applied stats we are welcoming *Devin Caughey, *a Professor of
Political Science at MIT. Devin will be presenting a joint project
entitled *Bayesian Population Interpolation and Lasso-Based Target
Selection in Survey Weighting. *The abstract is included below and can be
found on the website (here
<http://projects.iq.harvard.edu/applied.stats.workshop-gov3009/presentations…>).
As usual, we will meet in CGIS Knafel Room 354 at 12 pm and lunch will be
served.
I look forward to seeing you all there! Thanks!
-- Dana Higgins
*Title: *Bayesian Population Interpolation and Lasso-Based Target Selection
in Survey Weighting
*Abstract:* We propose solutions to two important problems that have
received relatively little attention in the field of survey weighting: the
construction of population targets in the face of irregularly missing data,
and the optimal selection of weighting targets from the set of possible
auxiliary variables. Our solution to the first problem relies on a dynamic
Bayesian population-interpolation model that allows subpopulation estimates
in a given year to be informed by data from other years. To address the
second, we formulate the problem of target selection as one of variable
subset selection, for which we propose a lasso-based solution. We
demonstrate the usefulness of these techniques by using them to generate
weights for quota-sampled opinion polls from the early days of survey
research. Given the declining response rates, rising use of non-probability
samples, and growth in potential sources of auxiliary information in
modern-day polling, these methods have wide potential application in
contemporary survey research as well. (co-authored with Mallory Wang)