Over the past years, I focused on changepoint detection via fast dynamical programming algorithms. Lately, I have been looking at online changepoint and anomaly detection algorithms. I am also interested in producing scalable techniques robust to those scenarios where the usual normality assumptions fall. Additionally, my interests span from modeling in general to other topics in data science, such as machine learning, and MCMC.
Publications
Published
Fast Online Changepoint Detection via Functional Pruning CUSUM statistics
G Romano, IA Eckley, P Fearnhead, G Rigaill - Journal of Machine Learning Research, 24, 1-36 (2023)gfpop: an R Package for Univariate Graph-Constrained Change-point Detection
V Runge, T D Hocking, G Romano, F Afghah, P Fearnhead, G Rigaill - Journal of Statistical Software, 106(6), 1-39 (2023)
Online non-parametric changepoint detection with application to monitoring operational performance of network devices
E Austin, G Romano, IA Eckley, P Fearnhead - Computational Statistics & Data Analysis 177 (2023)Detecting Abrupt Changes in the Presence of Local Fluctuations and Autocorrelated Noise
G Romano, G Rigaill, V Runge, P Fearnhead - Journal of the American Statistical Association 117.540 (2022)Changes in microbial utilisation and fate of soil carbon following the addition of different fractions of anaerobic digestate to soils
M Cattin, K Semple, M Stutter, G Romano, A Lag Brotons, C Parry, B Surridge - European Journal of Soil Science (2021)
A Constant-per-Iteration Likelihood Ratio Test for Online Changepoint Detection for Exponential Family Models
K Ward, G Romano, IA Eckley, P Fearnhead - Statistics and Computing, Volume 34 (2024)A Log-Linear Non-Parametric Online Changepoint Detection Algorithm based on Functional Pruning
G Romano, IA Eckley, P Fearnhead - IEEE Transactions on Signal Processing (2024)Pre-prints
Online Multivariate Changepoint Detection: Leveraging Links With Computational Geometry
L Pishchagina, G Romano, P Fearnhead, V Runge, G Rigaill (2023)Software
changepoint_online: A Collection of Methods for Online Changepoint Detection
A collection of online changepoint algorithms in Python. Includes Focus for the whole exponential family, NPFocus and finally MDFocus
FOCuS: Fast Online Changepoint Detection via Functional Pruning CUSUM statistics
FOCuS is an algorithm for detecting changes in mean in real-time. This is achieved by a fast-recursive update of the well known Page-CUSUM statistic.
NUNC: Online Non-parametric Changepoint Detection via Rolling Windows
NUNC is an real-time non-parametric changepoint detection method. The Algorithm seeks for a change in distribution within a rolling window.
DeCAFS: Detecting Changes in Autocorrelated and Fluctuating Signals
Detects abrupt changes in time series with local fluctuations as a random walk process and autocorrelated noise as an AR(1) process.
Conferences, Workshop and Events
Contributed
NeurIPS 2023, New Orleans (10-16 Dec 2023)
[Poster Session] FOCuS: Online Changepoint Detection via Functional Pruning CUSUM Statistics.
IMS Annual Meeting, London (27-30 June 2022)
FOCuS: Online Changepoint Detection via Functional Pruning CUSUM Statistics.
ISNPS2022, Paphos (20-24 June 2022)
NP-FOCuS: a Nonparametric Approach for Online Changepoint Detection.
EcoSta 2021 - Virtual Conference (24-26 Jun 2021)
Online changepoint detection with a constant per-iteration computational cost.
StatScale Workshop - Virtual Workshop (22-23 Apr 2021)
FOCuS: A CUSUM statistics for fast online changepoint detection.
CFE-CMStatistics - London (14-16 Dec 2019)
Detecting Abrupt Changes in Correlated Time-Series.
Attended
Changepoint and anomaly detection in big data settings - Paris
13-14 Nov 2019
APTS modules in Durham (8-12 Jul 2019)
Modules on Computer Intensive Statistics and on High-dimensional Statistics
APTS modules in Cambridge (10-14 Dec 2018)
Modules in Statistical Inference and Statistical Computing