Change sign detection with differential MDL change statistics and its applications to COVID-19 pandemic analysis

Yamanishi, Kenji; Xu, Linchuan; Yuki, Ryo; Fukushima, Shintaro; Lin, Chuan-hao

doi:10.1038/s41598-021-98781-4

Cited by 8 publications

(3 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The wave ends based on the same criteria as before. This methodology combines elements from incidence-based methods that detect infection waves by comparing the incidence rate of cases to a preset threshold ( 21 , 22 ), as well as change point detection methods which enable to detect waves that start meanwhile a previous wave is still decaying ( 23–25 ).…”

Section: Resultsmentioning

confidence: 99%

Predicting the spread of SARS-CoV-2 variants: An artificial intelligence enabled early detection

Levi,

Zerhouni,

Altuvia

2023

PNAS Nexus

View full text Add to dashboard Cite

During more than 3 years since its emergence, SARS-CoV-2 has shown great ability to mutate rapidly into diverse variants, some of which turned out to be very infectious and have spread throughout the world causing waves of infections. At this point, many countries have already experienced up to six waves of infections. Extensive academic work has focused on the development of models to predict the pandemic trajectory based on epidemiological data, but none has focused on predicting variant-specific spread. Moreover, important scientific literature analyzes the genetic evolution of SARS-CoV-2 variants and how it might functionally affect their infectivity. However, genetic attributes have not yet been incorporated into existing epidemiological modeling that aims to capture infection trajectory. Thus, this study leverages variant-specific genetic characteristics together with epidemiological information to systematically predict the future spread trajectory of newly detected variants. The study describes the analysis of 9.0 million SARS-CoV-2 genetic sequences in 30 countries and identifies temporal characteristic patterns of SARS-CoV-2 variants that caused significant infection waves. Using this descriptive analysis, a machine-learning-enabled risk assessment model has been developed to predict, as early as 1 week after their first detection, which variants are likely to constitute the new wave of infections in the following 3 months. The model’s out-of-sample area under the curve (AUC) is 86.3% for predictions after 1 week and 90.8% for predictions after 2 weeks. The methodology described in this paper could contribute more broadly to the development of improved predictive models for variants of other infectious viruses.

show abstract

Section: Resultsmentioning

confidence: 99%

Predicting the spread of SARS-CoV-2 variants: An artificial intelligence enabled early detection

Levi,

Zerhouni,

Altuvia

2023

PNAS Nexus

View full text Add to dashboard Cite

show abstract

“…Although they can be thought of as early warning signals of model changes, they cannot explain the intrinsic model dimensionality nor how rapidly a model changes in the transition period. Change sign detection method usng differential MDL change statistics has been proposed in [36]. However, it is applied to change sign detection for parameters only not for models.…”

Section: Related Workmentioning

confidence: 99%

Detecting Signs of Model Change with Continuous Model Selection Based on Descriptive Dimensionality

Yamanishi¹,

Hirai²

2023

Preprint

View full text Add to dashboard Cite

We address the issue of detecting changes of models that lie behind a data stream. The model refers to an integer-valued structural information such as the number of free parameters in a parametric model. Specifically we are concerned with the problem of how we can detect signs of model changes earlier than they are actualized. To this end, we employ continuous model selection on the basis of the notion of descriptive dimensionality (Ddim). It is a real-valued model dimensionality, which is designed for quantifying the model dimensionality in the model transition period. Continuous model selection is to determine the real-valued model dimensionality in terms of Ddim from a given data. We propose a novel methodology for detecting signs of model changes by tracking the rise-up of Ddim in a data stream. We apply this methodology to detecting signs of changes of the number of clusters in a Gaussian mixture model and those of the order in an auto regression model. With synthetic and real data sets, we empirically demonstrate its effectiveness by showing that it is able to visualize well how rapidly model dimensionality moves in the transition period and to raise early warning signals of model changes earlier than they are detected with existing methods.

show abstract

“…Furthermore, the issues of gradual changes have been discussed to investigate the transition periods for absolute changes. The MDL change statistics [ 32 ] and differential MDL change statistics [ 33 ] were proposed to measure the degree of gradual changes. The notions of structural entropy [ 34 ] and graph entropy [ 35 ] were proposed to measure the degree of model uncertainty in the changes.…”

Section: Related Workmentioning

confidence: 99%

Mixture Complexity and Its Application to Gradual Clustering Change Detection

Kyoya

Yamanishi

2022

Entropy

View full text Add to dashboard Cite

We consider measuring the number of clusters (cluster size) in the finite mixture models for interpreting their structures. Many existing information criteria have been applied for this issue by regarding it as the same as the number of mixture components (mixture size); however, this may not be valid in the presence of overlaps or weight biases. In this study, we argue that the cluster size should be measured as a continuous value and propose a new criterion called mixture complexity (MC) to formulate it. It is formally defined from the viewpoint of information theory and can be seen as a natural extension of the cluster size considering overlap and weight bias. Subsequently, we apply MC to the issue of gradual clustering change detection. Conventionally, clustering changes have been regarded as abrupt, induced by the changes in the mixture size or cluster size. Meanwhile, we consider the clustering changes to be gradual in terms of MC; it has the benefits of finding the changes earlier and discerning the significant and insignificant changes. We further demonstrate that the MC can be decomposed according to the hierarchical structures of the mixture models; it helps us to analyze the detail of substructures.

show abstract

Change sign detection with differential MDL change statistics and its applications to COVID-19 pandemic analysis

Cited by 8 publications

References 47 publications

Predicting the spread of SARS-CoV-2 variants: An artificial intelligence enabled early detection

Predicting the spread of SARS-CoV-2 variants: An artificial intelligence enabled early detection

Detecting Signs of Model Change with Continuous Model Selection Based on Descriptive Dimensionality

Mixture Complexity and Its Application to Gradual Clustering Change Detection

Contact Info

Product

Resources

About