How does the Journal webinar work?
Journal webinars are held every few months and last about an hour. Journal papers are carefully selected from recent issues of the
Royal Statistical Society's journals by editorial board for their importance, relevance and/or use of cutting-edge methodology; and authors are invited to present their work and take questions from an audience who 'dial in' to access the webinar.
Two papers are selected from our journals and authors will be invited to present their papers (20 minutes) followed by discussion (25 minutes) for each paper. Attendees will dial into a teleconference call. Papers and slides will be available to download two weeks in advance or you can log into the conference system and follow the presentation live online. These sessions are open to members and non-members. No need to pre-register. Audio recordings will be available for download shortly after the session.
Questions on the paper or general queries can be emailed in advance of the session to firstname.lastname@example.org.
For UK and international users the link to the Microsoft Teams app is provided on this page.
January 22 2020 at 3pm (UK time)
The webinar will feature two papers on online and offline methods of change point detection methods:
Paper 1: Online inference for multiple changepoint problems’ by Fearnhead & Liu.
It was published in Series B, Volume 69, Issue 4, September 2007
Presenter 1: Paul Fearnhead.
Abstract: We propose an on‐line algorithm for exact filtering of multiple changepoint problems. This algorithm enables simulation from the true joint posterior distribution of the number and position of the changepoints for a class of changepoint models. The computational cost of this exact algorithm is quadratic in the number of observations. We further show how resampling ideas from particle filters can be used to reduce the computational cost to linear in the number of observations, at the expense of introducing small errors, and we propose two new, optimum resampling algorithms for this problem. One, a version of rejection control, allows the particle filter to choose the number of particles that are required at each time step automatically. The new resampling algorithms substantially outperform standard resampling algorithms on examples that we consider; and we demonstrate how the resulting particle filter is practicable for segmentation of human G+C content.
Paper 2:High dimensional change point estimation via sparse projection’ by Wang & Samworth.
It was published in Series B, Volume 80, Issue1 in January 2018
Presenter 2: Tengyao Wang
Abstract: Change points are a very common feature of ‘big data’ that arrive in the form of a data stream. We study high dimensional time series in which, at certain time points, the mean structure changes in a sparse subset of the co‐ordinates. The challenge is to borrow strength across the co‐ordinates to detect smaller changes than could be observed in any individual component series. We propose a two‐stage procedure called inspect for estimation of the change points: first, we argue that a good projection direction can be obtained as the leading left singular vector of the matrix that solves a convex optimization problem derived from the cumulative sum transformation of the time series. We then apply an existing univariate change point estimation algorithm to the projected series. Our theory provides strong guarantees on both the number of estimated change points and the rates of convergence of their locations, and our numerical studies validate its highly competitive empirical performance for a wide range of data‐generating mechanisms. Software implementing the methodology is available in the R package InspectChangepoint.
Chair: Yi Yu, University of Warwick
Discussant: Claudia Kirch, Otto von Guericke University Magdeburg
The Wang & Samworth paper is open access. The Fearnhead & Liu paper will be available to download free of charge from two weeks before the webinar.
An open discussion led by our discussants will follow the author presentations in which everyone is encouraged to take part. You can ask the authors a question over the phone or type a message if you prefer using the web based teleconference system.
Questions can also be emailed in advance and further information requested from email@example.com.
Full event details including slides (when available) can be found on our website.
Journal webinars are free, open to everyone and simple to join.
Those unable to listen in live can listen to the podcast and view slides from the presentation afterwards on YouTube, accessible from the main RSS website.
Webcasts, MP3s and slides from past events are available to download.