Generate Time Series Research Articles

Toward coherent space–time mapping of seagrass cover from satellite data: an example of a Mediterranean lagoon

Abstract. Seagrass meadows are a highly productive and economically important shallow coastal habitat. Their sensitivity to natural and anthropogenic disturbances, combined with their importance for local biodiversity, carbon stocks, and sediment dynamics, motivate a frequent monitoring of their distribution. However, generating time series of seagrass cover from field observations is costly, and mapping methods based on remote sensing require restrictive conditions on seabed visibility, limiting the frequency of observations. In this contribution, we examine the effect of accounting for environmental factors, such as the bathymetry and median grain size (D50) of the substrate as well as the coordinates of known seagrass patches, on the performance of a random forest (RF) classifier used to determine seagrass cover. Using 148 Landsat images of the Venice Lagoon (Italy) between 1999 and 2020, we trained an RF classifier with only spectral features from Landsat images and seagrass surveys from 2002 and 2017. Then, by adding the features above and applying a time-based correction to predictions, we created multiple RF models with different feature combinations. We tested the quality of the resulting seagrass cover predictions from each model against field surveys, showing that bathymetry, D50, and coordinates of known patches exert an influence that is dependent on the training Landsat image and seagrass survey chosen. In models trained on a survey from 2017, where using only spectral features causes predictions to overestimate seagrass surface area, no significant change in model performance was observed. Conversely, in models trained on a survey from 2002, the addition of the out-of-image features and particularly coordinates of known vegetated patches greatly improves the predictive capacity of the model, while still allowing the detection of seagrass beds absent in the reference field survey. Applying a time-based correction eliminates small temporal variations in predictions, improving predictions that performed well before correction. We conclude that accounting for the coordinates of known seagrass patches, together with applying a time-based correction, has the most potential to produce reliable frequent predictions of seagrass cover. While this case study alone is insufficient to explain how geographic location information influences the classification process, we suggest that it is linked to the inherent spatial auto-correlation of seagrass meadow distribution. In the interest of improving remote-sensing classification and particularly to develop our capacity to map vegetation across time, we identify this phenomenon as warranting further research.

Understanding climate resilience in Scandinavia during the Neolithic and Early Bronze Age

Mid and late-Holocene climate shifts are considered to have profoundly shaped demographic developments and adaptive responses of communities globally. Yet their onset, duration, and impact on Neolithic and Early Nordic Bronze Age communities in the high-latitude ranges of southern and north-western Scandinavia remain a major research gap. Here, we built on an emerging body of archaeological and paleoclimate data, encompassing 20,908 anthropogenic 14C dates and 49 climate records from the Holocene. Additionally, we gathered and correlated a new archaeological dataset of 3649 houses from southern Scandinavia and southern Norway. In this study, we utilised 6268 reliable 14C dates and 2519 dwellings to generate time series and socio-economic trends from ∼4100 to 1100 BCE.Our study revealed three key findings: (1) A distinct lateral zonation, with variations in the duration and timing of the Holocene Thermal Maximum (∼7050–2050 BCE). In Southern Scandinavia, a warmer climate may have facilitated the spread of crop cultivation (3820–3790 BCE), coinciding with significant population growth. Neolithic communities settled in permanent two-aisled houses 90–160 years later (3700–3660 BCE). (2) The 2250 BCE (4.2 ka BP) cooling trend marked the beginning of a climate regime shift with varying duration and timing (∼3450–1450 BCE). This period coincided with demographic growth, migration, crop cultivation diversity, and the development of houses with crop storage facilities (2290–2215 BCE). (3) Severe abrupt cooling periods (∼1850–1450 BCE) corresponded to short-term demographic decline including disruptions in trade networks with continental Europe. However, repopulation and redistribution of wealth (∼1450 BCE), along with the development of stable three-aisled houses (1475–1450 BCE), underscore the resilience of food-producing economies in mitigating environmental disturbances.

A semi-parametric hourly space–time weather generator

Abstract. Long continuous time series of meteorological variables (i.e. rainfall, temperature and radiation) are required for applications such as derived flood frequency analyses. However, observed time series are generally too short, too sparse in space or incomplete, especially at the sub-daily timestep. Stochastic weather generators overcome this problem by generating time series of arbitrary length. This study presents a major revision to an existing space–time hourly rainfall model based on a point alternating renewal process, now coupled to a k-NN resampling model for conditioned simulation of non-rainfall climate variables. The point-based rainfall model is extended into space by the resampling of simulated rainfall events via a simulated annealing optimisation approach. This approach enforces observed spatial dependency as described by three bivariate spatial rainfall criteria. A new non-sequential branched shuffling approach is introduced which allows the modelling of large station networks (N>50) with no significant loss in the spatial dependence structure. Modelling of non-rainfall climate variables, i.e. temperature, humidity and radiation, is achieved using a non-parametric k-nearest neighbour (k-NN) resampling approach, coupled to the space–time rainfall model via the daily catchment rainfall state. As input, a gridded daily observational dataset (HYRAS) was used. A final deterministic disaggregation step was then performed on all non-rainfall climate variables to achieve an hourly output temporal resolution. The proposed weather generator was tested on 400 catchments of varying size (50–20 000 km2) across Germany, comprising 699 sub-daily rainfall recording stations. Results indicate no major loss of model performance with increasing catchment size and a generally good reproduction of observed climate and rainfall statistics.

Sig‐Wasserstein GANs for conditional time series generation

AbstractGenerative adversarial networks (GANs) have been extremely successful in generating samples, from seemingly high‐dimensional probability measures. However, these methods struggle to capture the temporal dependence of joint probability distributions induced by time‐series data. Furthermore, long time‐series data streams hugely increase the dimension of the target space, which may render generative modeling infeasible. To overcome these challenges, motivated by the autoregressive models in econometric, we are interested in the conditional distribution of future time series given the past information. We propose the generic conditional Sig‐WGAN framework by integrating Wasserstein‐GANs (WGANs) with mathematically principled and efficient path feature extraction called the signature of a path. The signature of a path is a graded sequence of statistics that provides a universal description for a stream of data, and its expected value characterizes the law of the time‐series model. In particular, we develop the conditional Sig‐ metric that captures the conditional joint law of time series models and use it as a discriminator. The signature feature space enables the explicit representation of the proposed discriminators, which alleviates the need for expensive training. We validate our method on both synthetic and empirical dataset and observe that our method consistently and significantly outperforms state‐of‐the‐art benchmarks with respect to measures of similarity and predictive ability.

Sample sizes for estimating the sensitivity of a monitoring system that generates repeated binary outcomes with autocorrelation.

Sample size formulas are provided to determine how many events and how many patient care units are needed to estimate the sensitivity of a monitoring system. The monitoring systems we consider generate time series binary data that are autocorrelated and clustered by patient care units. Our application of interest is an automated hand hygiene monitoring system that assesses whether healthcare workers perform hand hygiene when they should. We apply an autoregressive order 1 mixed effects logistic regression model to determine sample sizes that allow the sensitivity of the monitoring system to be estimated at a specified confidence level and margin of error. This model overcomes a major limitation of simpler approaches that fail to provide confidence intervals with the specified levels of confidence when the sensitivity of the monitoring system is above 90%.

TSGBench: Time Series Generation Benchmark

Synthetic Time Series Generation (TSG) is crucial in a range of applications, including data augmentation, anomaly detection, and privacy preservation. Although significant strides have been made in this field, existing methods exhibit three key limitations: (1) They often benchmark against similar model types, constraining a holistic view of performance capabilities. (2) The use of specialized synthetic and private datasets introduces biases and hampers generalizability. (3) Ambiguous evaluation measures, often tied to custom networks or downstream tasks, hinder consistent and fair comparison. To overcome these limitations, we introduce TSGBench, the inaugural Time Series Generation Benchmark, designed for a unified and comprehensive assessment of TSG methods. It comprises three modules: (1) a curated collection of publicly available, real-world datasets tailored for TSG, together with a standardized preprocessing pipeline; (2) a comprehensive evaluation measures suite including vanilla measures, new distance-based assessments, and visualization tools; (3) a pioneering generalization test rooted in Domain Adaptation (DA), compatible with all methods. We have conducted comprehensive experiments using TSGBench across a spectrum of ten real-world datasets from diverse domains, utilizing ten advanced TSG methods and twelve evaluation measures. The results highlight the reliability and efficacy of TSGBench in evaluating TSG methods. Crucially, TSGBench delivers a statistical analysis of the performance rankings of these methods, illuminating their varying performance across different datasets and measures and offering nuanced insights into the effectiveness of each method.

Statistical characterization and exploitation of Synthetic Aperture radar vegetation indexes for the generation of Leaf area Index time series

Statistical characterization and exploitation of Synthetic Aperture radar vegetation indexes for the generation of Leaf area Index time series

Evaluating stochastic rainfall models for hydrological modelling

Stochastic rainfall models are important tools for evaluating hydrological risks such as flooding and drought because of their ability to randomly generate alternative plausible climatic timeseries. The stochastic generation of climatic timeseries is not an end in itself, since they are typically applied to a catchment to determine the performance of water-related infrastructure systems, such as reservoirs or flood-control measures. This methodology typically involves a train of models to determine the end-of-system impact, yet the evaluation of stochastically generated rainfall timeseries is usually a stand-alone procedure focused on metrics directly related to the stochastic generator. This paper demonstrates discrepancies in this approach by evaluating two, daily-timestep, stochastic rainfall models in terms of rainfall metrics and their subsequently generated flow metrics after rainfall-runoff transformation. The two models are a Markov-based model and a latent-variable model, where each model is calibrated and evaluated showing ‘overall good’ performance. Stochastically generated timeseries, alongside observed rainfall timeseries are inputted to a calibrated catchment model (GR4J) to derive daily flow timeseries. Whereas the rainfall metrics typically showed ‘good’ performance, streamflow-based metrics are not necessarily ‘good’. The procedure is repeated for 277 stations from Australia and 106 stations from the United States of America. Depending on the strictness of the flow-based comparison and region analysed, using the Markov-based model 12–26% of sites were classified as ‘poor’ performing, and 1%-9% of sites were classified as ‘poor’ using the latent-variable model. The results demonstrate that catchment-based performance of flow metrics is more holistic since it magnifies features of the rainfall not otherwise visible to rainfall-based evaluation.

A study on pronunciation assessment of English learners based on temporal classification algorithm

Abstract This paper utilizes a time-series classification algorithm to classify samples, and the selected datasets are utilized to calculate the distance of time-series samples. The information gain is calculated using the information obtained, and the information entropy is determined collectively for the node datasets to generate time series features. The convolution operation is used to obtain the formal representation of the time series classification, and the extracted English pronunciation features are adaptively matched. The evaluation results were determined using a multilayer wavelet feature scale transformation method, and English learners achieved scores of 6 and above in all three tests using the time-series classification algorithm. To master standard English pronunciation, English learners should use the temporal classification algorithm.

A model-free-based control chart for batch process using U-statistics

Batch processes are known to generate time series of successive measurements of many process variables in each run and the main challenge is to capture and accommodate the variability in the batch domain (batch-to-batch variability) and in the time domain (data with serial correlation). Classical approaches are focused on the first goal by applying multivariate techniques in the columns of a data matrix, in which each row has the data for the entire batch run. Those approaches are grounded in the traditional Multiway Principal Component Analysis (MPCA). Recent approaches are focused on the balance between the two sources, taking into account the time series nature of the data mainly by using time series models and neural network methods to directly accommodate the variability in the time domain. This paper proposes a model-free approach based on U-statistics theory. Through this theory, we can build a group of control charts (named V charts) capable of monitoring batch processes considering both sources of variability. In general, the new batch under monitoring is evaluated according to a group of reference in-control samples, based on the distance measures from those batches. The proposed approach is really flexible since we avoid model identification, parameter estimates issues and it is able to accommodate the within and between batch variability. Additionally, our method is suitable to deal with a wide range of batch data features, since under weak conditions, the distribution of the V statistics remains well known. It makes the V-based charts able to be used in different kinds of processes. We are focused here on controlling changes caused by data drifts (including trends) and by data serial correlation (data dynamics). Through the simulated and real batch process, we show that the V charts work well, including a scenario with very few reference batches available.

Production Improvement Rate with Time Series Data on Standard Time at Manufacturing Sites

Amid the changes brought about by the 4th Industrial Revolution, numerous studies have been undertaken to develop smart factories, with a strong emphasis on knowledge-based manufacturing through smart factory construction. Advances in manufacturing data collection, fusion, and mining technologies have significantly bolstered the utilization of knowledge-based manufacturing. Data mining technology is widely employed for facility maintenance and failure prediction. Smart factory operations are pursuing automation and autonomization. Automation of production planning is also essential to achieve automation and autonomy in factory operations, from planning to execution. With the advancement of data mining technology, it is possible to automate production planning for the production planning and prediction of future production through information based on current conditions based on the past. The baseline information generated based on the current situation is suitable for automating short-term operational planning. If we generate time series reference information based on data from the past to the present, we can also automate long-term operation planning. By measuring the results of productivity improvements in mass-produced products from the past to the present and extrapolating them to future products, time series baseline information on production time is generated. If the baseline information is used for long-term planning, it can be used to predict future production capacity and facility shortages. This study presents a methodology and utilization method for calculating the rate of change in production time, which can be applied to production plan prediction and equipment investment capacity forecasting in future factory operations, using historical time series production time data.

Solar-Mixer: An Efficient End-to-End Model for Long-Sequence Photovoltaic Power Generation Time Series Forecasting

The expansion of photovoltaic power generation makes photovoltaic power forecasting an essential requirement. With the development of deep learning, more accurate predictions have become possible. This paper proposes an efficient end-to-end model for solar power generation that allows for long-sequence time series forecasting.This paper proposes an efficient end-to-end system for solar power generation that allows for long-term forecasting. Two modules comprise the forecasting model: the anomaly detection module and the forecasting module. Singular values are detected and corrected by the anomaly detection module. And in the forecasting module, the series is first divided into different intervals. After that, the divided data is passed through a decomposition module embedded in the neural network, and each interval is encoded as high-dimension vectors. These vectors are fed into the mixing layer, which is used to learn the relationship between intervals and channels after the encoding process. Every mixing layer comprises two different layers: the channel-mixing layer allows communication between different channels, and the interval-mixing layer allows communication between different intervals. This neural network is purely dependent on the multilayer perceptron, resulting in a network with low system latency and low training cost. The experiments results not only show the model proposed in this paper can make accurate long-sequence time series forecasts for photovoltaic power generation at three sites but also show that the model can defeat the state-of-the-art model in long-sequence photovoltaic power generation time series forecasting work.

Missing interpolation model for wind power data based on the improved CEEMDAN method and generative adversarial interpolation network

Randomness and fluctuations in wind power output may cause changes in important parameters (e.g., grid frequency and voltage), which in turn affect the stable operation of a power system. However, owing to external factors (such as weather), there are often various anomalies in wind power data, such as missing numerical values and unreasonable data. This significantly affects the accuracy of wind power generation predictions and operational decisions. Therefore, developing and applying reliable wind power interpolation methods is important for promoting the sustainable development of the wind power industry. In this study, the causes of abnormal data in wind power generation were first analyzed from a practical perspective. Second, an improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) method with a generative adversarial interpolation network (GAIN) network was proposed to preprocess wind power generation and interpolate missing wind power generation sub-components. Finally, a complete wind power generation time series was reconstructed. Compared to traditional methods, the proposed ICEEMDAN-GAIN combination interpolation model has a higher interpolation accuracy and can effectively reduce the error impact caused by wind power generation sequence fluctuations

A new index for the rapid generation of chlorophyll time series from hyperspectral imaging of sediment cores

AbstractTime series analyses of pigment concentrations are key to understanding past aquatic ecosystem dynamics. As lake sediments provide a window into longer‐term changes, innovative paleolimnological chlorophyll quantification could provide impactful insights into past environmental processes. Lab‐based hyperspectral imaging of sediment cores is an emerging technique to develop rapid, non‐destructive, high‐resolution chlorophyll inferences but it requires more extensive vetting. Despite recent advances in model development, there is still a knowledge gap about the reliability of chlorophyll models when applied in lakes with diverse properties, as well as the potential confounding effects of physical sediment properties on these models. We assessed the performance of 23 chlorophyll indices based on paired measurements collected via hyperspectral imaging and spectrophotometry for 202 samples spread across seven Canadian lake sediment cores. The best performance was by a new index based on the wavelength of the red‐edge minimum point (λREMP). We tested the applicability of λREMP to a broad range of sediment cores using a database of 116 cores, and found the index to provide reliable reconstructions of ƩChl (i.e., chlorophyll a and b and their degradation products) trends in 84% of sites. Further analyses indicated that sediment characteristics including particle size, organic matter content, water content, and density had no systematic impact on ƩChl, but greater sediment brightness did increase ƩChl inferences from hyperspectral images. Hyperspectral core scanning is poised to facilitate the generation of high‐resolution chlorophyll time series data, which could greatly improve our understanding of trajectories of change from the local to global scales.

Framework of wind joint analysis for different lake regions and its effects on the water quality

Water quality of Lake Taihu (LT) is paid wide attention, and the wind field makes wind-induced flow, significantly impacting the water quality. As a wide lake, the wind field suffers spatial and temporal changes. The correlated relationship among the wind fields for different lake regions has not been studied, and the joint effects of wind field on the water quality is still unknown. Hence, this paper proposed a framework of wind joint analysis for different lake regions and modelling its effects on the water quality, including joint analysis of wind field for different lake regions using Copula function, random time series generation of wind field using Monte-Carlo simulation, and hydrodynamic and water quality numerical simulation. Taking the Lake Taihu (LT) as study case, the joint relationships of wind field for four lake regions were analyzed, and the CODMN, TP and TN distributions under different wind fields were simulated. This paper showed the following: (1) The relationship between the wind speed and direction was weak, and the correlated relationships among the wind fields for four lake regions were significant (ρ > 0.5). (2) The water quality of LT was more influenced by the wind direction rather than the wind speed, with an impacting ratio of ∼4.5. (3) The discharged weighted angle of runoff outflow (DWARO) was a practical index to improve the water quality. When regulating the input/output runoff discharge, water diversion project and water intakes in practice, the administrations should make the DWARO less than or equal to the average wind direction, to decrease the average pollutant concentration and improve the water quality of lake and intake. The proposed framework and results of LT could be extended to other wide shallow lakes to support the environmental management and basin runoff regulation.

Disparities in preterm birth following the July 1995 Chicago heat wave

PurposeTo evaluate if changes in preterm birth (PTB, <37 weeks of gestation) incidence differed between non-Hispanic (NH) Black and NH white births following the July 1995 Chicago heat wave—among the most severe U.S. heat waves since 1950. MethodsWe used an ecologic study design. We obtained birth data from January 1990–December 1996 from the National Vital Statistics File to calculate the mean monthly PTB incidence in Chicago’s Cook County, Illinois. Births between July 1995 and February 1996 were potentially exposed to the heat wave in utero. We generated time series models for NH Black and NH white births, which incorporated synthetic controls of Cook County based on unexposed counties. We ran a secondary analysis considering socioeconomic status (SES). ResultsFrom 1990–1996, the mean monthly PTB incidence among NH Black births was 18.6% compared to 7.8% among NH white births. The mean monthly PTB incidence among NH Black births from August 1995–January 1996 was 16.7% higher than expected (three additional PTBs per 100 live births per month [95% confidence interval (CI): 1, 5]). A similar increase occurred among low-SES NH Black births. No increase appeared among NH white births. ConclusionsSevere heat waves may increase racial disparities in PTB incidence.

Understanding of the properties of neural network approaches for transient light curve approximations

Context. Modern-day time-domain photometric surveys collect a lot of observations of various astronomical objects and the coming era of large-scale surveys will provide even more information on their properties. Spectroscopic follow-ups are especially crucial for transients such as supernovae and most of these objects have not been subject to such studies. Aims. Flux time series are actively used as an affordable alternative for photometric classification and characterization, for instance, peak identifications and luminosity decline estimations. However, the collected time series are multidimensional and irregularly sampled, while also containing outliers and without any well-defined systematic uncertainties. This paper presents a search for the best-performing methods to approximate the observed light curves over time and wavelength for the purpose of generating time series with regular time steps in each passband. Methods. We examined several light curve approximation methods based on neural networks such as multilayer perceptrons, Bayesian neural networks, and normalizing flows to approximate observations of a single light curve. Test datasets include simulated PLAsTiCC and real Zwicky Transient Facility Bright Transient Survey light curves of transients. Results. The tests demonstrate that even just a few observations are enough to fit the networks and improve the quality of approximation, compared to state-of-the-art models. The methods described in this work have a low computational complexity and are significantly faster than Gaussian processes. Additionally, we analyzed the performance of the approximation techniques from the perspective of further peak identification and transients classification. The study results have been released in an open and user-friendly Fulu Python library available on GitHub for the scientific community.

ANALYSIS OF AEROSOLS IN WEST AFRICA: MODELLING AND RADIATIVE FORCING

We studied the aerosol properties in Cinzana and Ilorin in West Africa using data of aerosols’ properties over the period 2000–2015 from MODIS (Moderate resolution Imaging Spectroradiometer). There has been limited report on the comparison between this two West African Stations. The time series modelling, prediction of aerosol parameters and the estimation of aerosol radiative forcing (ARF) were focused on. Statistical Package for the Social Sciences (SPSS) software was used for the modelling the monthly averaged measurements of aerosol optical depth at 550 nm (τ550), Angstrom exponent at 470 – 660 nm (α470-660), cloud fraction (Ncloud), fine mode fraction (FMF) and single scattering albedo at 470 nm (ωo) data over the two stations while the ARF values were calculated. The generated time series models are all good fit and the values are biased, both negatively and positively. The ARF of the two stations’ aerosol data shows value between 0.0 and -750.0 W/m2 for Ilorin and -10.0 to -200. 0 W/m2, which results mainly in cooling. Although there is a difference between ARF values in the two stations, both stations demonstrated clear seasonal variation with the rainy season having the least negative aerosol forcing. The results suggest that the aerosols are predominantly mineral dust, but those in Cinzana have more absorbing properties than those in Ilorin. Overall, the research shows that the modelled parameters can be utilized in lieu of the measured data and that the solar radiation received in this region has been attenuated at various degrees

A Bio-Inspired Chaos Sensor Model Based on the Perceptron Neural Network: Machine Learning Concept and Application for Computational Neuro-Science.

The study presents a bio-inspired chaos sensor model based on the perceptron neural network for the estimation of entropy of spike train in neurodynamic systems. After training, the sensor on perceptron, having 50 neurons in the hidden layer and 1 neuron at the output, approximates the fuzzy entropy of a short time series with high accuracy, with a determination coefficient of R2~0.9. The Hindmarsh-Rose spike model was used to generate time series of spike intervals, and datasets for training and testing the perceptron. The selection of the hyperparameters of the perceptron model and the estimation of the sensor accuracy were performed using the K-block cross-validation method. Even for a hidden layer with one neuron, the model approximates the fuzzy entropy with good results and the metric R2~0.5 ÷ 0.8. In a simplified model with one neuron and equal weights in the first layer, the principle of approximation is based on the linear transformation of the average value of the time series into the entropy value. An example of using the chaos sensor on spike train of action potential recordings from the L5 dorsal rootlet of rat is provided. The bio-inspired chaos sensor model based on an ensemble of neurons is able to dynamically track the chaotic behavior of a spike signal and transmit this information to other parts of the neurodynamic model for further processing. The study will be useful for specialists in the field of computational neuroscience, and also to create humanoid and animal robots, and bio-robots with limited resources.

Generative Adversarial Networks applied to synthetic financial scenarios generation

In this paper, we introduce Jinkou, a GAN-based algorithm that allows for the conditional generation of synthetic multivariate time series. The set of variables whose distribution is to be replicated include specific variables taking different values for different objects, as well state variables describing the state of the world, common to all objects at a given date and potentially influential on the specific features. The conditioning process is specified at inference time, and only involves state variables; it simply consists in setting lower and/or upper bounds on their values. The generative model is trained as an un-conditioned generator and is agnostic of any scenario the user might set at inference time. The use case considered in this pilot study is of interest for the financial industry: the generator produces random samples of the instrument-specific features over time (e.g their price, size or the risk for securities). Such generation is conditioned on user-defined macroeconomic assumptions/scenarios involving global variables, such as inflation, oil prices or interest rates. We introduce numerical metrics to assess the statistical closeness between the two multivariate distributions of historical and artificial data. As proof of concept, we test the proposed algorithm by reproducing the value variation for two possible portfolios, Energy and Financial, conditioned on scenarios for which a consensus is present in the community. Jinkou allows us to recover some classical stylized facts about the financial markets, this ability constituting a proof of its efficiency.