Detecting multiple generalized change-points by isolating single ones

A Selective Review on Information Criteria in Multiple Change Point Detection

Change points indicate significant shifts in the statistical properties in data streams at some time points. Detecting change points efficiently and effectively are essential for us to understand the underlying data-generating mechanism in modern data streams with versatile parameter-varying patterns. However, it becomes a highly challenging problem to locate multiple change points in the noisy data. Although the Bayesian information criterion has been proven to be an effective way of selecting multiple change points in an asymptotical sense, its finite sample performance could be deficient. In this article, we have reviewed a list of information criterion-based methods for multiple change point detection, including Akaike information criterion, Bayesian information criterion, minimum description length, and their variants, with the emphasis on their practical applications. Simulation studies are conducted to investigate the actual performance of different information criteria in detecting multiple change points with possible model mis-specification for the practitioners. A case study on the SCADA signals of wind turbines is conducted to demonstrate the actual change point detection power of different information criteria. Finally, some key challenges in the development and application of multiple change point detection are presented for future research work.

A Multiple change-point detection framework on linguistic characteristics of real versus fake news articles

Extracting information from textual data of news articles has been proven to be significant in developing efficient fake news detection systems. Pointedly, to fight disinformation, researchers concentrated on extracting information which focuses on exploiting linguistic characteristics that are common in fake news and can aid in detecting false content automatically. Even though these approaches were proven to have high performance, the research community proved that both the language as well as the word use in literature are evolving. Therefore, the objective of this paper is to explore the linguistic characteristics of fake news and real ones over time. To achieve this, we establish a large dataset containing linguistic characteristics of various articles over the years. In addition, we introduce a novel framework where the articles are classified in specified topics based on their content and the most informative linguistic features are extracted using dimensionality reduction methods. Eventually, the framework detects the changes of the extracted linguistic features on real and fake news articles over the time incorporating a novel change-point detection method. By employing our framework for the established dataset, we noticed that the linguistic characteristics which concern the article's title seem to be significantly important in capturing important movements in the similarity level of "Fake" and "Real" articles.

Detection and estimation of multiple transient changes

Change-point detection methods are proposed for the case of temporary failures, or transient changes, when an unexpected disorder is ultimately followed by a re-adjustment and return to the initial state. A base distribution of the 'in-control' state changes to an 'out-of-control' distribution for unknown periods of time. Likelihood based sequential and retrospective tools are proposed for the detection and estimation of each pair of change-points. The accuracy of the obtained change-point estimates is assessed. Proposed methods offer simultaneous control of the familywise false alarm and false re-adjustment rates at the pre-chosen levels.

The impact of government actions and risk perception on the promotion of self-protective behaviors during the COVID-19 pandemic

INTRODUCTION

We aim to understand the factors that drive citizens of different countries to adhere to recommended self-protective behaviors during the COVID-19 pandemic.

METHODS

Survey data was obtained through the COVID-19 Impact project. We selected countries that presented a sufficiently complete time series and a statistically relevant sample for running the analysis: Cyprus, Germany, Greece, Ireland, Latvia, Spain, Switzerland, the United Kingdom, and the United States of America. To identify country-specific differences in self-protective behaviors, we used previous evidence and change-point detection analysis to establish variations across participating countries whose effect was then assessed by means of interrupted series analysis.

RESULTS

A high level of compliance with health and governmental authorities' recommendations were generally observed in all included countries. The level of stress decreased near the period when countries such as Cyprus, Greece or the United Kingdom relaxed their prevention behavior recommendations. However, this relaxation of behaviors did not occur in countries such as Germany, Ireland, or the United States. As observed in the change-point detection analysis, when the daily number of recorded COVID-19 cases decreased, people relaxed their protective behaviors (Cyprus, Greece, Ireland), although the opposite trend was observed in Switzerland.

DISCUSSION

COVID-19 self-protective behaviors were heterogeneous across countries examined. Our findings show that there is probably no single winning strategy for exiting future health crises, as similar interventions, aimed to promote self-protective behaviors, may be received differently depending on the specific population groups and on the particular geographical context in which they are implemented.

Testing for a Change in Mean After Changepoint Detection

While many methods are available to detect structural changes in a time series, few procedures are available to quantify the uncertainty of these estimates post-detection. In this work, we fill this gap by proposing a new framework to test the null hypothesis that there is no change in mean around an estimated changepoint. We further show that it is possible to efficiently carry out this framework in the case of changepoints estimated by binary segmentation and its variants, ℓ 0 segmentation, or the fused lasso. Our setup allows us to condition on much less information than existing approaches, which yields higher powered tests. We apply our proposals in a simulation study and on a dataset of chromosomal guanine-cytosine content. These approaches are freely available in the R package ChangepointInference at https://jewellsean.github.io/changepoint-inference/.