Was that change real? Quantifying uncertainty for change points
Lead Research Organisation:
Lancaster University
Department Name: Mathematics and Statistics
Abstract
Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.
Organisations
Publications
Carrington Rachel
(2023)
Improving Power by Conditioning on Less in Post-selection Inference for Changepoints
in arXiv e-prints
Cho H
(2023)
High-Dimensional Time Series Segmentation via Factor-Adjusted Vector Autoregressive Modeling
in Journal of the American Statistical Association
Li Jie
(2022)
Automatic Change-Point Detection in Time Series via Deep Learning
in arXiv e-prints
| Description | The grant has made advances in two areas. The first is developing ideas at the interface of changepoint detection and machine learning/AI: with the first set of ideas for how to automatically create change-point detection methods based on training a deep neural network on examples (either real or simulated) of data with and without a change. The second is improving the power of methods that assign p-values, a measure of certainty, to detected change-points. |
| Exploitation Route | The work on using AI to create new approaches to change-point detection is one of the first papers in this area, and it opens up a range of opportunities to develop this general idea across many change-point problems. This would be of interest to academic researchers in statistics, signal processing, computer science, machine learning and AI. The ability to quantify certainty of detected changes is important across a range of applications -- with initial work in this area motivated by neuroscientists wishing to quantify uncertainty around detected spikes in calcium imaging data. |
| Sectors | Aerospace Defence and Marine Healthcare Pharmaceuticals and Medical Biotechnology |
| Title | Rchangepoint |
| Description | Code for calculating p-values for changes detected by a range of change detection algorithms. This implements the methods in Improving Power by Conditioning on Less in Post-selection Inference for Changepoints Rachel Carrington and Paul Fearnhead |
| Type Of Technology | Software |
| Year Produced | 2023 |
| Impact | The software enables reproduction of results in the paper "Improving Power by Conditioning on Less in Post-selection Inference for Changepoints". |
| URL | https://arxiv.org/pdf/2301.05636.pdf |