Change Points In Time Series Research Articles

Trend analysis and change-point detection of temperature and rainfall in southern Peruvian Amazon and its relation to deforestation

The study aimed to identify the change points, tendencies, and trends in climatic parameters (precipitation and temperatures) and to investigate their relationship with deforestation in the southeastern Peruvian Amazon (Tambopata). Rainfall and temperature data for the Puerto Maldonado station from 1970 to 2023 was used. Monthly, seasonal, and annual precipitation as well as temperature (maximum, minimum, and mean) were analyzed for possible trends using nonparametric Mann-Kendal statistic test, while the Pettitt test was employed to detect the abrupt change point in time series. The Spearman's correlation coefficient was used to identify the relationship between deforestation and climate parameters. The results revealed a rise in mean, minimum, and maximum temperatures. Mann Kendall and Sen’s slope revealed significant trends in the monthly, seasonal and annual temperatures in the study period. However, in contrast to the temperature variation trend, the monthly, seasonal and annual precipitation did not present a significant trend. Significant positive correlations were obtained between deforestation and temperatures but its association with precipitation was not significant.

Multi-Sensors Time Series Change Point Detection in Wireless Sensor Networks Based on Deep Evidential Fusion and Self Distillation Learning

Multi-Sensors Time Series Change Point Detection in Wireless Sensor Networks Based on Deep Evidential Fusion and Self Distillation Learning

ОБНАРУЖЕНИЕ РАЗЛАДОК В НЕСТАЦИОНАРНЫХ РЯДАХ НАБЛЮДЕНИЙ: ОБЗОР ТЕХНОЛОГИЙ И КРИТИЧЕСКИЙ АНАЛИЗ

A critical analysis of modern mathematical technologies used to solve the problem of nonstationary time series change point, taking into account the chaotic nature of the observed processes, is considered. The known methods of the quickest change point detection in nonstationary time series described by traditional models of random processes are considered as initial assumptions. The discrepancy between the known time series observation models and the real processes occurring in unstable immersion environments described by the concept of stochastic chaos is shown. Time series segmentation technologies based on a posteriori analysis of changes in the properties of observation sites of chaotic processes are presented. Options for building online algorithms for the quickest change point detection based on methods of multiregressive data analysis and metric machine learning technologies are considered.

Trend Analysis and Change Points in Time Series of Water Level of Shallow and Deep Wells in Gorganrood Watershed

Trend Analysis and Change Points in Time Series of Water Level of Shallow and Deep Wells in Gorganrood Watershed

Research on time series change point detection and influencing factors under machine learning: based on PM2.5 concentration data in Hefei city

Research on time series change point detection and influencing factors under machine learning: based on PM2.5 concentration data in Hefei city

A Least Squares Estimator for Gradual Change-Point in Time Series with m-Asymptotically Almost Negatively Associated Errors

As a new member of the NA (negative associated) family, the m-AANA (m-asymptotically almost negatively associated) sequence has many statistical properties that have not been developed. This paper mainly studies its properties in the gradual change point model. Firstly, we propose a least squares type change point estimator, then derive the convergence rates and consistency of the estimator, and provide the limit distributions of the estimator. It is interesting that the convergence rates of the estimator are the same as that of the change point estimator for independent identically distributed observations. Finally, the effectiveness of the estimator in limited samples can be verified through several sets of simulation experiments and an actual hydrological example.

Influencing factors of PM2.5 concentration in the typical urban agglomerations in China based on wavelet perspective

PM2.5 is one of the most harmful air pollutants affecting sustainable economic and social development in China. The analysis of influencing factors affecting PM2.5 concentration is significant for the improvement of air quality. In this study, three typical urban agglomerations in China (Beijing‒Tianjin‒Hebei [BTH], the Yangtze River Delta [YRD], and the Pearl River Delta [PRD]) were studied using innovative trend analysis, a Bayesian statistical model, and partial wavelet and multiwavelet coherence to analyze PM2.5 concentration variations and multi-scale coupled oscillations between PM2.5 concentration and air pollutants/meteorological factors. The results showed that: (1) PM2.5 concentration time-series showed significant downward trends, which decreased as follows: BTH > YRD > PRD. The higher the pollution level, the greater the change trend. In BTH and the PRD, PM2.5 had obvious trends and seasonal change points; whereas, the PM2.5 time-series change point in the YRD was not obvious. (2) PM2.5 had significant intermittent resonance cycles with air pollutants and meteorological factors in different time domains. There were differences in the main controlling factors affecting PM2.5 among the three urban agglomerations. (3) The explanatory ability of air pollutant combinations for variations in PM2.5 was higher than that of meteorological factor combinations. However, the synergistic effect of air pollutants/meteorological factors could better explain the PM2.5 concentration variations on all time-frequency scales. The results of this study provide a reference for ecological improvement as well as collaborative governance of air pollution.

Assessing the impacts of reservoirs on downstream hydrological frequency based on a general rainfall-reservoir index

When large reservoirs are built and put into operation, the downstream hydrological processes will be altered significantly, and ecology and agricultural irrigation water of the basin will be affected to some extent. The reservoir index (RI) and the sediment trapping efficiency (TE) of reservoirs are defined to quantify the reservoir impacts on the water flow and sediment by considering the static storage capacity. However, the regulating effect of reservoirs on hydrological variables is not only related to static storage capacity, but also to dynamic reservoir operation. Thus, in this paper, a general rainfall-reservoir index (GRRI) is developed by coupling reservoir regulation indicator (RR, including RI and TE) and effective rainfall affecting the dynamic operation of reservoirs, and the GRRI is used as the covariate to carry out the nonstationary frequency analysis of flood (Q) and annual sediment load (S) at Gaochang (GC) station in Min River, Wulong (WL) station in Wu River, Ankang (AK), Huangjiagang (HJG) and Huangzhuang (HZ) station in Han River, and Cuntan (CT) station on the main stream of the upper Yangtze River. It is found that Q and S at six stations have obvious changes induced by reservoirs, the mean of Q decreases by 22.8%–60.6%, and S drops by 47.7%–89.5% after the change-point of time series. The nonstationary probability distribution models with GRRI as the covariate have better fitting effects than nonstationary models with RR as the covariate. With the incorporation of the impacts of effective rainfall, the GRRI can more accurately capture the occurrence of nonstationarity in the downstream hydrological frequency. These results might be helpful for exploring the impact mechanism of the reservoir regulation on the downstream hydrological variables as well as ecological management of basin.

Improving Detection of Changepoints in Short and Noisy Time Series with Local Correlations: Connecting the Events in Pixel Neighbourhoods

AbstractDetecting changepoints in time series becomes difficult when the series are short and the observation variance is high. In the context of time series of environmental resource maps, it is often safe to assume that the abrupt events are spatially continuous, and so are the changepoints. We propose to utilise this assumption by means of hierarchical models where the changepoints are modelled using a spatial model. We demonstrate utility of the approach by constructing a Bayesian model based on the Potts model, with additional assumptions relevant to changepoint detection in national multi-source forest inventory maps. We discuss implementation issues and demonstrate the idea’s performance using a simulation study. We then apply the model to forest resource maps in order to detect felling events.

Changing trends of patient characteristics and treatment pathways during the COVID-19 pandemic: A cross-sectional analysis of 72,459 inpatient cases from the German Helios database.

This study compared patient profiles and clinical courses of SARS-CoV-2 infected inpatients over different pandemic periods. In a retrospective cross-sectional analysis, we examined administrative data of German Helios hospitals using ICD-10-codes at discharge. Inpatient cases with SARS-CoV-2 infection admitted between 03/04/2020 and 07/19/2022 were included irrespective of the reason for hospitalization. All endpoints were timely assigned to admission date for trend analysis. The first pandemic wave was defined by change points in time-series of incident daily infections and compared with different later pandemic phases according to virus type predominance. We included 72,459 inpatient cases. Patients hospitalized during the first pandemic wave (03/04/2020-05/05/2020; n = 1,803) were older (68.5 ± 17.2 vs. 64.4 ± 22.6 years, p < 0.01) and severe acute respiratory infections were more prevalent (85.2 vs. 53.3%, p < 0.01). No differences were observed with respect to distribution of sex, but comorbidity burden was higher in the first pandemic wave. The risk of receiving intensive care therapy was reduced in all later pandemic phases as was in-hospital mortality when compared to the first pandemic wave. Trend analysis revealed declines of mean age and Elixhauser comorbidity index over time as well as a decline of the utilization of intensive care therapy, mechanical ventilation and in-hospital mortality. Characteristics and outcomes of inpatients with SARS-CoV-2 infection changed throughout the observational period. An ongoing evaluation of trends and care pathways will allow for the assessment of future demands.

Applying Grey Relational Analysis to Detect Change Points in Time Series

The goal of detecting change points is to recognize abrupt changes in time series data. This is suitable, for instance, to find events that characterize the financial market or to inspect data streams of stock returns. Regression models categorized as supervised methods have played a significant role in change‐point detection. However, since change points might not be available beforehand to train the model, and because the series data might be statistically atypical, the applicability of regression models is limited. To avoid statistical assumptions, this study uses the grey theory, a kind of artificial intelligence tools, to measure the relationships between sequences by grey relational analysis (GRA). This paper contributes to propose an unsupervised method to detect possible change points in time series by GRA. Change‐point analysis of the proposed method was performed on S&amp;P100 stock returns. Experimental results from evaluating the recognition accuracy rate show that the proposed method performs well compared to other methods considered for change‐point detection.

Bayesian Online Change Point Detection in Finance

Abstract It is quite common that the structure of a time series changes abruptly. Identifying these change points and describing the model structure in the segments between these change points is an important task in financial time series analysis. Change point detection is the identification of abrupt changes in the generative parameters of sequential data. In application areas such as finance, online rather than offline detection of change points in time series is mostly required, due to their use in predictive tasks, possibly embedded in automatic trading systems. However, the complex structure of the data generation processes makes this a challenging endeavor. This paper is concerned with online change point detection in financial time series using the Bayesian setting. To this end, the Bayesian posterior probability of change at a specific time is proposed and some procedures are presented for selecting the priors and estimation of parameters. Applications in simulated financial time series are given. Finally, conclusions are proposed.

744. Incidence and Duration of Diagnostic Delays Associated with Dengue Fever

Abstract Background Dengue fever is a prominent emerging arboviral infection in the tropics and subtropics, and an important cause of systemic febrile illness among some international travelers. Signs and symptoms are similar to more common infectious illnesses in temperate climates, and dengue may not be promptly considered when patients seek evaluation. Methods We conducted a retrospective cohort study of patients diagnosed with dengue fever using the IBM MarketScan Research database from 2001-2017. We identified cases of dengue fever where patients were enrolled ≥ 1 year prior to the index diagnosis. All healthcare visits in the year prior to the index diagnosis were collected and we identified visits with signs/symptoms compatible with dengue or a diagnosis made of an illness with similar symptoms (e.g., influenza) before the index dengue diagnosis. We used a time-series change-point analysis to identify the time before diagnosis in which symptoms of dengue became more prominent. We conducted a bootstrap-based simulation analysis to estimate the duration and frequency of missed diagnostic opportunities. Results We identified 4,449 cases of dengue fever that met eligibility criteria. We found that 2,791 (62.7%) had ≥ 1 healthcare visit(s) prior to diagnosis with characteristic symptoms of dengue recorded. Our simulations analysis supports that 32.9% (95% CI: 31.1-35.0) experienced 1 or more missed opportunities for diagnosis. Among these patients, the average duration of diagnostic delay was 8.26 (CI: 6.32-11.38) days and ~21% of patients had a diagnostic delay of 2 or more weeks. Patients with a delayed diagnosis averaged 2.2 (CI 2.11-2.29) healthcare visits which represented missed opportunities. Missed opportunities were more likely during weekend, ED or outpatient visits. Conclusion Dengue fever is not considered in the majority of patients at the time of the initial symptomatic evaluation in the U.S., indicating delays in diagnosis are common. Enhanced education of providers about dengue fever could lead to more prompt diagnosis that should help optimize patient management. Disclosures All Authors: No reported disclosures

CPMAN: Change Point Detection Approach in Time Series Based on the Prediction of Multi-stage Attention Networks

Time series change point detection can identify the locations of abrupt points in many dynamic processes. It can help us to find anomaly data in an early stage. At the same time, detecting change points for long, periodic, and multiple input series data has received a lot of attention recently, and is universally applicable in many fields including power, environment, finance, and medicine. However, the performance of classical methods typically scales poorly for such time series. In this paper, we propose CPMAN, a novel prediction-based change point detection approach via multi-stage attention networks. Our model consists of two key modules. First, in the time series prediction module, we employ the multi-stage attention-based networks and integrate the multi-series fusion mechanism. This module can adaptively extract features from the relevant input series and capture the long-term temporal dependencies. Secondly, in the change point detection module, we use the wavelet analysis-based algorithm to detect change points efficiently and identify the change points and outliers. Extensive experiments are conducted on various real-world datasets and synthetic datasets, proving the superiority and effectiveness of CPMAN. Our approach outperforms the state-of-the-art methods by up to 12.1% on the F1 Score.

Change Point Analysis for Detecting Vaccine Safety Signals.

It is important to detect signals of abrupt changes in adverse event reporting in order to notice public safety concerns and take prompt action, especially for vaccines under national immunization programs. In this study, we assessed the applicability of change point analysis (CPA) for signal detection in vaccine safety surveillance. The performances of three CPA methods, namely Bayesian change point analysis, Taylor’s change point analysis (Taylor-CPA), and environmental time series change point detection (EnvCpt), were assessed via simulated data with assumptions for the baseline number of events and degrees of change. The analysis was validated using the Korea Adverse Event Reporting System (KAERS) database. In the simulation study, the Taylor-CPA method exhibited better results for the detection of a change point (accuracy of 96% to 100%, sensitivity of 7% to 100%, specificity of 98% to 100%, positive predictive value of 25% to 85%, negative predictive value of 96% to 100%, and balanced accuracy of 53% to 100%) than the other two CPA methods. When the CPA methods were applied to reports of syncope or dizziness following human papillomavirus (HPV) immunization in the KAERS database, Taylor-CPA and EnvCpt detected a change point (Q2/2013), which was consistent with actual public safety concerns. CPA can be applied as an efficient tool for the early detection of vaccine safety signals.

Accounting for multiple testing in the analysis of spatio-temporal environmental data

The statistical analysis of environmental data from remote sensing and Earth system simulations often entails the analysis of gridded spatio-temporal data, with a hypothesis test being performed for each grid cell. When the whole image or a set of grid cells are analyzed for a global effect, the problem of multiple testing arises. When no global effect is present, we expect alpha % of all grid cells to be false positives, and spatially autocorrelated data can give rise to clustered spurious rejections that can be misleading in an analysis of spatial patterns. In this work, we review standard solutions for the multiple testing problem and apply them to spatio-temporal environmental data. These solutions are independent of the test statistic, and any test statistic can be used (e.g., tests for trends or change points in time series). Additionally, we introduce permutation methods and show that they have more statistical power. Real-world data are used to provide examples of the analysis, and the performance of each method is assessed in a simulation study. Unlike other simulation studies, our study compares the statistical power of the presented methods in a comprehensive simulation study. In conclusion, we present several statistically rigorous methods for analyzing spatio-temporal environmental data and controlling the false positives. These methods allow the use of any test statistic in a wide range of applications in environmental sciences and remote sensing.

An Adaptive Intrusion Detection Scheme for Cloud Computing

To provide dynamic resource management, live virtual machine migration is used to move a virtual machine from one host to another. However, virtual machine migration poses challenges to cloud intrusion detection systems because movement of VMs from one host to another makes it difficult to create a consistent normal profile for anomaly detection. Hence, there is a need to provide an adaptive anomaly detection system capable of adapting to changes that occur in the cloud data during VM migration. To achieve this, the authors proposed a scheme for adaptive IDS for Cloud computing. The proposed adaptive scheme is comprised of four components: an ant colony optimization-based feature selection component, a statistical time series change point detection component, adaptive classification, and model update component, and a detection component. The proposed adaptive scheme was evaluated using simulated datasets collected from vSphere and performance comparison shows improved performance over existing techniques.

Analysis of hydrometeorological variables over the transboundary Komadugu-Yobe basin, West Africa

Abstract Spatiotemporal trends in daily observed precipitation, river discharge, maximum and minimum temperature data were investigated between 1971 and 2013 in the Komadugu-Yobe basin. Significant change points in time series are corrected using Adapted Caussinus-Mestre Algorithm for homogenizing Networks of Temperature series algorithm. Mann–Kendall test and Sen's slope are used to estimate the trend and its magnitude at dry, wet and annual season time scales, respectively. Preliminary results show an increasing trend of the observed variables. There is a latitudinal increase (decrease) in the basin temperature (precipitation) from lower to higher latitudes. The minimum temperature (0.05 °C/year) increases faster than the maximum temperature (0.03 °C/year). Overall, the percentage changes in minimum temperature range between 3 and 10% while that of maximum temperature ranges between 1 and 3%. Due to precipitation dependence on regional characteristics, the highest percentage change was recorded in precipitation with values between −5 and 97%. In all time scales, river discharge and precipitation have strong positive correlations while the correlation between river discharge and temperature is negative. It is imperative to advocate and support positive developmental practices as well as establishing necessary mitigation measures to cope with the effects of climate in the basin.

Detection of Change Points in Time Series with Moving Average Filters and Wavelet Transform: Application to EEG Signals

We investigated change point detection (CPD) in time series composed of harmonic functions driven by Gaussian noise (in EEGs, in particular) and proposed a method of moving average filters in conjunction with wavelet transform. Numerical simulations showed that CPD runs over 90% within the frequency band <40 Hz. This means that detection of structural change points is almost guaranteed in the respective cases. The mean absolute error (MAE) as a measure of performance of the method was below 5%. The method is rather robust against noise. It has been demonstrated that CPD is possible at the noise amplitude exceeding 25% of the amplitude of harmonic functions. In application of the proposed method on the signals, CPD appeared in 74% of the analyzed EEGs.

How detecting change-points in biomarkers time series can support the diagnostic approaches of ovarian cancer

Ovarian cancer is the fifth most common cause of cancer-related deaths among women, with more than 150,000 deceases worldwide each year. One of the main problems of ovarian cancer is that, usually, is detected at later stages, for which the survival rates are relatively low. Thus, the development of new approaches for longitudinal multi-marker analysis that result in earlier detection may significantly impact on mortality. An issue of specific interest is to determine whether the baseline level of a biomarker changes significantly at some time instant (change-point) prior to the clinical diagnosis of cancer. Here we apply a hierarchical Bayesian change-point model to jointly study the features of time series from multiple biomarkers, such as the serum biomarker Cancer Antigen 125, the Human Epididymis Protein 4, matrix metalloproteinase-7, Cytokeratin 19 fragment, glycodelin and mesothelin, using data from a nested case-control study of women diagnosed with ovarian cancer in the UK Collaborative Trial of Ovarian Cancer Screening. In this way we assess whether some of these biomarkers can play a role in change-point detection and, therefore, aid in the diagnosis of the disease. The main conclusion of our study is that the combined analysis of a group of specific biomarkers may possibly improve the detection of changepoints in time series data (compared to the analysis of Cancer Antigen 125 alone, the most commonly used oncomarker in the screening of ovarian cancer) which, in turn, are relevant for the early diagnosis of ovarian cancer.