Predicting the Swedish Election Using a Bayesian State Space Model (Botten Ada)
A talk about the undelying Bayesian model that powered “Botten Ada”
Abstract
In the Swedish election 2014 “Botten Ada”, a website using Bayesian modelling and probabilistic machine learning made predictions on the election outcome. In the upcoming Swedish election in the fall, “Botten Ada” and it’s founders, Måns Magnusson, Assistant Professor in statistics at Uppsala University and J++ a data journalism team based in Stockholm, will make a comeback with a new updated and refined model together with Jonas Wallin, Lund University. What degree of accuracy can be reached? And what place can prediction models such as these have in political science, and in academia at large?
Citation
BibTeX citation:
@unpublished{wallin2022,
author = {Wallin, Jonas and Magnusson, Måns and Finnäs, Jens},
title = {Predicting the {Swedish} {Election} {Using} a {Bayesian}
{State} {Space} {Model} {(Botten} {Ada)}},
date = {2022-06-08},
url = {https://www.iffs.se/},
langid = {en},
abstract = {In the Swedish election 2014 “Botten Ada”, a website using
Bayesian modelling and probabilistic machine learning made
predictions on the election outcome. In the upcoming Swedish
election in the fall, “Botten Ada” and it’s founders, Måns
Magnusson, Assistant Professor in statistics at Uppsala University
and J++ a data journalism team based in Stockholm, will make a
comeback with a new updated and refined model together with Jonas
Wallin, Lund University. What degree of accuracy can be reached? And
what place can prediction models such as these have in political
science, and in academia at large?}
}
For attribution, please cite this work as:
Wallin, Jonas, Måns Magnusson, and Jens Finnäs. 2022. “Predicting
the Swedish Election Using a Bayesian State Space Model (Botten
Ada).” Institutet för framtidsstudier, Stockholm, June 8. https://www.iffs.se/.