Article
Statistics & Probability
Bertille Follain, Tengyao Wang, Richard J. Samworth
Summary: The proposed method introduces a new way to estimate changepoints in partially observed, high-dimensional time series that undergo a simultaneous change in mean. It demonstrates strong effectiveness and efficiency in both simulated data and real oceanographic dataset analysis.
JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-STATISTICAL METHODOLOGY
(2022)
Review
Meteorology & Atmospheric Sciences
Robert B. Lund, Claudie Beaulieu, Rebecca Killick, Qiqi Lu, Xueheng Shi
Summary: This paper reviews the methods used to identify and analyze the changepoints in climate data, with a focus on helping scientists make reliable conclusions. The paper discusses common mistakes and pitfalls to avoid in changepoint analysis and provides recommendations for best practices. The paper also provides examples of how these methods have been applied to temperature and sea ice data. The main goal of the paper is to provide guidance on how to effectively identify the changepoints in climate time series and homogenize the series.
JOURNAL OF CLIMATE
(2023)
Article
Economics
Zhaoxing Gao, Ruey S. Tsay
Summary: This paper introduces a new procedure for building factor models for high-dimensional unit-root time series, which involves a non-singular linear transformation, estimation of stationary common factors, and idiosyncratic white noise components. The proposed method is shown to have good performance in terms of forecasting ability for a 508-dimensional PM2.5 series in Taiwan, compared to other commonly used methods.
INTERNATIONAL JOURNAL OF FORECASTING
(2021)
Article
Economics
Kashif Yousuf, Serena Ng
Summary: High dimensional predictive regressions are useful in a wide range of applications, but theories assuming stationary models with time invariant parameters may not align with evidence of parameter instability in economic time series. This paper introduces two L-2 boosting algorithms for estimating high dimensional models with smoothly evolving coefficients and locally stationary predictors. The study establishes the consistency of both methods and explores the benefits of modeling time variation in macroeconomic forecasting.
JOURNAL OF ECONOMETRICS
(2021)
Article
Economics
Bin Chen, Kenwin Maung
Summary: In this paper, a new nonparametric estimator is proposed for time-varying forecast combination weights. The theoretical properties and empirical performance of the estimator are demonstrated through the study of local linear estimation and penalized local linear estimation.
JOURNAL OF ECONOMETRICS
(2023)
Article
Computer Science, Artificial Intelligence
Ruobin Gao, Okan Duru, Kum Fai Yuen
Summary: This paper introduces a new method to address the lag-selection problem by utilizing supervised principal component analysis to reduce the lag structure to a lower dimensional space, overcoming practicality issues with grid search in high dimensions and randomness in evolutionary algorithms.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Statistics & Probability
Shaojun Guo, Xinghao Qiao
Summary: In this paper, a three-step procedure is proposed to estimate high-dimensional functional time series models. A novel functional stability measure based on spectral properties is introduced, providing theoretical guarantees for the procedure. The non-asymptotic properties of relevant estimated terms are also investigated. Additionally, a regularization approach is developed to estimate autoregressive coefficient functions under the sparsity constraint, and the finite-sample performance of the proposed method is evaluated through simulations and a public financial dataset.
Article
Statistics & Probability
Rong Chen, Dan Yang, Cun-Hui Zhang
Summary: This article introduces a factor model approach for analyzing high-dimensional dynamic tensor time series and multi-category dynamic transport networks, presenting two estimation procedures, their theoretical properties, and simulation results. Two applications are provided to illustrate the model and its interpretations.
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
(2022)
Article
Statistics & Probability
Bo Zhang, Guangming Pan, Qiwei Yao, Wang Zhou
Summary: We propose a new unsupervised learning method for clustering time series based on a latent factor structure, where each cluster has its own specific factors and common factors that impact all the time series. The estimation of the common factors, cluster-specific factors, and latent clusters is shown to have explicit convergence rates. Numerical illustrations using simulated and real data are provided, and the proposed approach also advances statistical inference for the factor model of Lam and Yao.
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
(2023)
Article
Statistics & Probability
Likai Chen, Weining Wang, Wei Biao Wu
Summary: This paper introduces a method for detecting multiple change-points in high-dimensional time series, supported by theoretical consistency and asymptotic distribution. The proposed two-step procedure can capture both the biggest break across different coordinates and aggregating simultaneous breaks over multiple coordinates.
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
(2022)
Article
Thermodynamics
Hugo Vinicius Bitencourt, Luiz Augusto Facury de Souza, Matheus Cascalho dos Santos, Rodrigo Silva, Petronio Candido de Lima e Silva, Frederico Gadelha Guimaraes
Summary: Given the use of multi-sensor environments and the two way communication between energy consumers and the smart grid, high-dimensional time series are increasingly arising in the Internet of Energy (IoE). Fuzzy Time Series (FTS) models, as data-driven non-parametric models of easy implementation and high accuracy, are of great value in smart building and IoE applications. However, existing FTS models can become unfeasible if all variables were used to train the model. Therefore, we propose a data-driven approach named Embedding Fuzzy Time Series (EFTS), which combines data embedding transformation and FTS methods, and shows superior accuracy and parsimony compared to baseline methods and previous literature results.
Article
Statistics & Probability
Bin Luo, Xiaoli Gao
Summary: This paper presents a framework for estimation in high-dimensional regression models using Penalized Robust Approximated quadratic M-estimators (PRAM). It addresses the issues of asymmetry, heteroscedasticity, and contamination that often occur with the growth of data dimensionality. Theoretical analysis shows that PRAM has local estimation consistency and oracle property. Computational results demonstrate satisfactory performance of PRAM under different irregular settings.
JOURNAL OF MULTIVARIATE ANALYSIS
(2022)
Article
Computer Science, Artificial Intelligence
Bo Chen, Min Fang, Xiao Li
Summary: Matrix factorization method has gained popularity in handling high-dimensional time series data. However, challenges still exist in long-term dependency management. To address this, we propose a novel approach that incorporates a latent bias effect and denoising model, improving the accuracy and robustness of the model.
NEURAL COMPUTING & APPLICATIONS
(2023)
Article
Mathematics
Mohammad Arashi, Mina Norouzirad, Mahdi Roozbeh, Naushad Mamode Khan
Summary: This paper introduces an estimation procedure for high-dimensional multicollinear data and investigates its asymptotic performance as the dimension grows. The proposed estimator is proven to be consistent and its asymptotic properties are derived. Monte Carlo simulation experiments are conducted to assess its performance, with a focus on analyzing a high-dimensional genetic dataset.
Article
Geosciences, Multidisciplinary
Kun Zhang, Wasif Bin Mamoon, E. Schwartz, Anthony. J. J. Parolari
Summary: Monitoring water quality at high frequency is difficult and expensive. Compressed sensing (CS) can be used to reconstruct high-frequency water quality data using limited measurements, as water quality signals are often sparse in the frequency domain. In this study, the sparsity of stream flow and concentration time-series was investigated, and CS was tested for reconstruction. CS effectively reconstructed the signals with only 5%-10% of the measurements needed. The study also found that CS can be integrated with dimensionality reduction and optimization techniques for more efficient sampling schemes in environmental geosciences and engineering.
GEOPHYSICAL RESEARCH LETTERS
(2023)
Article
Computer Science, Theory & Methods
J. D. B. Nelson, A. J. Gibberd, C. Nafornita, N. Kingsbury
STATISTICS AND COMPUTING
(2018)
Article
Statistics & Probability
A. Gibberd, S. Roy
Summary: This study examines the consistency properties of a regularized estimator for identifying both changepoints and graphical dependency structure in multivariate time-series. The Group-Fused Graphical Lasso (GFGL) is studied, which penalizes partial correlations with an L1 penalty while inducing block-wise smoothness over time to detect multiple changepoints. Consistency of the estimator in terms of both changepoints and the structure of graphical models in each segment is proven, contrasting with previous dynamic graph estimation methods conducted at a node-wise level.
ANNALS OF THE INSTITUTE OF STATISTICAL MATHEMATICS
(2021)
Article
Multidisciplinary Sciences
Ricardo Pio Monti, Alex Gibberd, Sandipan Roy, Matthew Nunes, Romy Lorenz, Robert Leech, Takeshi Ogawa, Motoaki Kawanabe, Aapo Hyvarinen
Article
Biology
E. A. K. Cohen, A. J. Gibberd
Summary: Wavelets provide flexibility for analyzing stochastic processes at different scales. In this article, we apply wavelets to multivariate point processes to detect and analyze unknown nonstationarity. We develop a temporally smoothed wavelet periodogram to ensure statistical tractability and demonstrate its equivalence to a multi-wavelet periodogram. Under the assumption of stationarity, the distribution of the temporally smoothed wavelet periodogram is shown to be asymptotically Wishart, with readily computable parameters. The distributional results also extend to wavelet coherence, a measure of inter-process correlation. We apply this statistical framework to construct a test for stationarity in multivariate point processes and successfully detect and characterize time-varying dependency patterns in neural spike-train data.
Article
Economics
Konstantinos Ferentinos, Alex Gibberd, Benjamin Guin
Summary: Climate policies aimed at reducing greenhouse gas emissions can lead to decreased asset values, risking them to become stranded. In this study, we examine the price effects of a specific climate policy, a minimum energy efficiency standard, in the housing market. Using a unique dataset of all residential house transactions in England and Wales, we find that energy-inefficient properties affected by this policy experienced an average decrease in prices by about £5,000 to £9,000 relative to efficient ones. We interpret this evidence as consistent with semi-strong market efficiency in the housing market.
Proceedings Paper
Computer Science, Information Systems
Alex J. Gibberd, Edward A. K. Cohen
2018 CONFERENCE RECORD OF 52ND ASILOMAR CONFERENCE ON SIGNALS, SYSTEMS, AND COMPUTERS
(2018)
Proceedings Paper
Engineering, Electrical & Electronic
J. D. B. Nelson, A. J. Gibberd
2016 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP)
(2016)
Proceedings Paper
Computer Science, Artificial Intelligence
Alex J. Gibberd, James D. B. Nelson
2015 IEEE GLOBAL CONFERENCE ON SIGNAL AND INFORMATION PROCESSING (GLOBALSIP)
(2015)