scholarly journals Identifying stationary phases in multivariate time-series for highlighting behavioural modes and home range settlements

2018 ◽  
Author(s):  
Rémi Patin ◽  
Marie-Pierre Étienne ◽  
Émilie Lebarbier ◽  
Simon Chamaillé-Jammes ◽  
Simon Benhamou
Keyword(s):  
Time Series ◽  
Home Range ◽  
Large Scale ◽  
Stationary Phases ◽  
Activity Level ◽  
Range Shifts ◽  
Small Scale ◽  
Minimum Length ◽  
List Type ◽  
Ancillary Data

AbstractRecent advances in bio-logging open promising perspectives in the study animal movements at numerous scales. It is now possible to record time-series of animal locations and ancillary data (e.g. activity level derived from on-board accelerometers) over extended areas and long durations with a high spatial and temporal resolution. Such time-series are often piecewise stationary, as the animal may alternate between different stationary phases (i.e. characterised by a specific mean and variance of some key parameter for limited periods). Identifying when these phases start and end is a critical first step to understand the dynamics of the underlying movement processes.We introduce a new segmentation-clustering method we called segclust2d. It can segment bi-(or more generally multi-) variate time-series and possibly cluster the various segments obtained, corresponding to phases assumed to be stationary. It is easy to use, as it only requires specifying the minimum length of a segment (to prevent over-segmentation) based on biological considerations.Although this method can be applied to time-series of any nature, we focus here on two-dimensional piecewise time-series whose phases correspond at small scale to the expressions of different behavioural modes such as transit, feeding and resting, as characterised by two joint metrics such as speed and turning angles or, at larger scale, to temporary home ranges, characterised by stationary distributions of bivariate coordinates.Using computer simulations, we show that segcust2d can rival and even outperform previous, more complex methods, which were specifically developed to highlight changes in movement modes or home range shifts (based on Hidden Markov or Ornstein-Uhlenbeck modelling, respectively), which, contrary to our method, require truly informative initial guesses to be efficient. Furthermore we demonstrate it on actual examples involving a zebra’s small scale movements and an elephant’s large scale movements, to illustrate the identification of various movement modes and of home range shifts, respectively.

scholarly journals Self-Supervised Pre-Training of Transformers for Satellite Image Time Series Classification

2020 ◽  
Author(s):  
Yuan Yuan ◽  
Lei Lin
Keyword(s):  
Time Series ◽  
Deep Learning ◽  
Large Scale ◽  
Temporal Structure ◽  
Satellite Image ◽  
Fine Tuning ◽  
Small Scale ◽  
Model Parameters ◽  
Learning Approaches ◽  
Wide Range

Satellite image time series (SITS) classification is a major research topic in remote sensing and is relevant for a wide range of applications. Deep learning approaches have been commonly employed for SITS classification and have provided state-of-the-art performance. However, deep learning methods suffer from overfitting when labeled data is scarce. To address this problem, we propose a novel self-supervised pre-training scheme to initialize a Transformer-based network by utilizing large-scale unlabeled data. In detail, the model is asked to predict randomly contaminated observations given an entire time series of a pixel. The main idea of our proposal is to leverage the inherent temporal structure of satellite time series to learn general-purpose spectral-temporal representations related to land cover semantics. Once pre-training is completed, the pre-trained network can be further adapted to various SITS classification tasks by fine-tuning all the model parameters on small-scale task-related labeled data. In this way, the general knowledge and representations about SITS can be transferred to a label-scarce task, thereby improving the generalization performance of the model as well as reducing the risk of overfitting. Comprehensive experiments have been carried out on three benchmark datasets over large study areas. Experimental results demonstrate the effectiveness of the proposed method, leading to a classification accuracy increment up to 1.91% to 6.69%. <div><b>This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible.</b></div>

scholarly journals Identifying stationary phases in multivariate time series for highlighting behavioural modes and home range settlements

2019 ◽  
Vol 89 (1) ◽  
pp. 44-56 ◽  
Author(s):  
Rémi Patin ◽  
Marie‐Pierre Etienne ◽  
Emilie Lebarbier ◽  
Simon Chamaillé‐Jammes ◽  
Simon Benhamou
Keyword(s):  
Time Series ◽  
Home Range ◽  
Stationary Phases ◽  
Multivariate Time Series

scholarly journals Self-Supervised Pre-Training of Transformers for Satellite Image Time Series Classification

2020 ◽  
Author(s):  
Yuan Yuan ◽  
Lei Lin
Keyword(s):  
Time Series ◽  
Deep Learning ◽  
Large Scale ◽  
Temporal Structure ◽  
Satellite Image ◽  
Fine Tuning ◽  
Small Scale ◽  
Model Parameters ◽  
Learning Approaches ◽  
Wide Range

<div>Satellite image time series (SITS) classification is a major research topic in remote sensing and is relevant for a wide range of applications. Deep learning approaches have been commonly employed for SITS classification and have provided state-of-the-art performance. However, deep learning methods suffer from overfitting when labeled data is scarce. To address this problem, we propose a novel self-supervised pre-training scheme to initialize a Transformer-based network by utilizing large-scale unlabeled data. In detail, the model is asked to predict randomly contaminated observations given an entire time series of a pixel. The main idea of our proposal is to leverage the inherent temporal structure of satellite time series to learn general-purpose spectral-temporal representations related to land cover semantics. Once pre-training is completed, the pre-trained network can be further adapted to various SITS classification tasks by fine-tuning all the model parameters on small-scale task-related labeled data. In this way, the general knowledge and representations about SITS can be transferred to a label-scarce task, thereby improving the generalization performance of the model as well as reducing the risk of overfitting. Comprehensive experiments have been carried out on three benchmark datasets over large study areas. Experimental results demonstrate the effectiveness of the proposed method, leading to a classification accuracy increment up to 2.38% to 5.27%. The code and the pre-trained model will be available at https://github.com/linlei1214/SITS-BERT upon publication.</div><div><b>This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible.</b></div>

scholarly journals Identifying Ecosystem Function Shifts in Africa Using Breakpoint Analysis of Long-Term NDVI and RUE Data

2020 ◽  
Vol 12 (11) ◽  
pp. 1894
Author(s):  
Thomas P. Higginbottom ◽  
Elias Symeonakis
Keyword(s):  
Time Series ◽  
Large Scale ◽  
Vegetation Index ◽  
Change Point Analysis ◽  
Annual Rainfall ◽  
Small Scale ◽  
Growth Season ◽  
Continental Scale ◽  
Normalised Difference Vegetation Index

Time-series of vegetation greenness data, derived from Earth-observation imagery, have become a key source of information for studying large-scale environmental change. The ever increasing length of such series allows for a range of indicators to be derived and for increasingly complex analyses to be applied. This study presents an analysis of trends in vegetation productivity—measured using the Global Inventory Monitoring and Modelling System third generation (GIMMS3g) Normalised Difference Vegetation Index (NDVI) data—for African savannahs, over the 1982–2015 period. Two annual metrics were derived from the 34 year dataset: the monthly, smoothed NDVI (the aggregated growth season NDVI) and the associated Rain Use Efficiency (growth season NDVI divided by annual rainfall). These indicators were then used in a BFAST-based change-point analysis, allowing the direction of change over time to change and the detection of one major break in the time-series. We also analysed the role of land cover type and climate zone as associations of the observed changes. Both methods agree that vegetation greening was pervasive across African savannahs, although RUE displayed less significant changes than NDVI. Monotonically increasing trends were the most common trend type for both indicators. The continental scale of the greening may suggest global processes as key drivers, such as carbon fertilization. That NDVI trends were more dynamic than RUE suggests that a large component of vegetation trends is driven by precipitation variability. Areas of negative trends were conspicuous by their minimalism. However, some patterns were apparent. In the southern Sahel and West Africa, declining NDVI and RUE overlapped with intensive population and agricultural regions. Dynamic trend reversals, in RUE and NDVI, located in Angola, Zambia and Tanzania, coincide with areas where a long-term trend of forest degradation and agricultural expansion has recently given way to increases in woody biomass. Meanwhile in southern Africa, monotonic increases in RUE with varying NDVI trend types may be indicative of shrub encroachment. However, all these processes are small-scale relative to the GIMMS NDVI data, and reconciling these conflicting drivers is not a trivial task. Our study highlights the importance of considering multiple options when undertaking trend analyses, as different inputs and methods can reveal divergent patterns.

scholarly journals Self-Supervised Pre-Training of Transformers for Satellite Image Time Series Classification

2020 ◽  
Author(s):  
Yuan Yuan ◽  
Lei Lin
Keyword(s):  
Time Series ◽  
Deep Learning ◽  
Large Scale ◽  
Temporal Structure ◽  
Satellite Image ◽  
Fine Tuning ◽  
Small Scale ◽  
Model Parameters ◽  
Learning Approaches ◽  
Wide Range

<div>Satellite image time series (SITS) classification is a major research topic in remote sensing and is relevant for a wide range of applications. Deep learning approaches have been commonly employed for SITS classification and have provided state-of-the-art performance. However, deep learning methods suffer from overfitting when labeled data is scarce. To address this problem, we propose a novel self-supervised pre-training scheme to initialize a Transformer-based network by utilizing large-scale unlabeled data. In detail, the model is asked to predict randomly contaminated observations given an entire time series of a pixel. The main idea of our proposal is to leverage the inherent temporal structure of satellite time series to learn general-purpose spectral-temporal representations related to land cover semantics. Once pre-training is completed, the pre-trained network can be further adapted to various SITS classification tasks by fine-tuning all the model parameters on small-scale task-related labeled data. In this way, the general knowledge and representations about SITS can be transferred to a label-scarce task, thereby improving the generalization performance of the model as well as reducing the risk of overfitting. Comprehensive experiments have been carried out on three benchmark datasets over large study areas. Experimental results demonstrate the effectiveness of the proposed method, leading to a classification accuracy increment up to 2.38% to 5.27%. The code and the pre-trained model will be available at https://github.com/linlei1214/SITS-BERT upon publication.</div><div><b>This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible.</b></div>

scholarly journals Multiscale estimation of the field-aligned current density

2018 ◽  
Author(s):  
Costel Bunescu ◽  
Joachim Vogt ◽  
Adrian Blagau ◽  
Octav Marghitu
Keyword(s):  
Time Series ◽  
Large Scale ◽  
Present Report ◽  
Synthetic Data ◽  
Minimum Variance ◽  
Small Scale ◽  
Spatial And Temporal Scales ◽  
Sampling Schemes ◽  
Multi Scale ◽  
Coupling Processes

Abstract. Field-aligned currents (FACs) in the magnetosphere-ionosphere (M-I) system exhibit a range of spatial and temporal scales that are linked to key dynamic coupling processes. To disentangle the scale dependence in magnetic field signatures of auroral FACs and to characterize their geometry and orientation, Bunescu et al. (2015) introduced the multi-scale FAC analyzer framework based on minimum variance analysis (MVA) of magnetic time series segments. In the present report this approach is carried further to include in the analysis framework a FAC density scalogram, i.e., a multiscale representation of the FAC density time series. The new technique is validated and illustrated using synthetic data consisting of overlapping sheets of FACs at different scales. The method is applied to Swarm data showing both large-scale and quiet aurora as well as mesoscale FAC structures observed during more disturbed conditions. We show both planar and non-planar FAC structures as well as uniform and non-uniform FAC density structures. For both, synthetic and Swarm data, the multiscale analysis is applied by two scale sampling schemes, namely the linear, and the logarithmic scanning of the FACs scale domain. The scale integrated FAC density is computed by both small-scale and large-scale weighting. The integrated multiscale FAC density is compared with the input FAC density for the synthetic data, whereas for the Swarm data we cross-check the results with well established single- and dual-spacecraft techniques. The entire multiscale information provides a new visualization tool for the complex FAC signatures, that complement other FAC analysis tools.

scholarly journals Tardive dyskinesia update: the syndrome

2018 ◽  
Vol 25 (1) ◽  
pp. 57-69 ◽  
Author(s):  
David Cunningham Owens
Keyword(s):  
Risk Factors ◽  
Tardive Dyskinesia ◽  
Large Scale ◽  
New Drugs ◽  
Psychiatric Research ◽  
Small Scale ◽  
List Type ◽  
Short Term ◽  
Prevalence Studies ◽  
Clinical Overview

SUMMARYTardive dyskinesia is a common iatrogenic neurological and neurobehavioural syndrome associated with the use of antidopaminergic medication, especially antipsychotics. Prior to the introduction of the newer antipsychotics in the 1990s, it was one of the major areas of psychiatric research but interest waned as the new drugs were reputed to have a reduced liability to extrapyramidal adverse effects in general, a claim now discredited by numerous pragmatic research studies. Early small-scale short-term prevalence studies were presented as evidence to support the assumption that patients on the newer drugs did indeed have a lower prevalence of tardive dyskinesia but recent large-scale review of studies with patients exposed for longer suggest that things have not changed. This article presents a clinical overview of a complex and varied syndrome in terms of its phenomenology, epidemiology and risk factors; a companion article will consider treatment. This overview aims to highlight tardive dyskinesia once again, especially to practitioners who have trained in an environment where this was considered mainly in historical terms.LEARNING OBJECTIVES•Understand the complex phenomenology comprising the syndrome of tardive dyskinesia•Appreciate recent data on prevalence and incidence with the newer antipsychotics•Be aware of risk factors when recommending antipsychotic (and other antidopaminergic) drugsDECLARATION OF INTERESTNone.

scholarly journals Daily to annual net primary production in the North Sea determined using autonomous underwater gliders and satellite Earth observation

2021 ◽  
Author(s):  
Benjamin Loveday ◽  
Timothy Smyth ◽  
Anıl Akpinar ◽  
Tom Hull ◽  
Mark Inall ◽  
...  
Keyword(s):  
Time Series ◽  
Primary Production ◽  
North Sea ◽  
Large Scale ◽  
Net Primary Production ◽  
Earth Observation ◽  
Small Scale ◽  
The North ◽  
The North Sea

Abstract. Shelf-seas play a key role in both the global carbon cycle and coastal marine ecosystems through the drawn-down and fixing of carbon, as measured through phytoplankton net primary production (NPP). Measuring NPP in situ, and extrapolating this to the local, regional and global scale presents challenges however because of limitations with the techniques utilised (e.g. radiocarbon isotopes), data sparsity and the inherent biogeochemical heterogeneity of coastal and open-shelf waters. Here, we introduce a powerful new technique based on the synergistic use of in situ glider profiles and satellite Earth Observation measurements which can be implemented in a real-time or delayed mode system. We apply this system to a fleet of gliders successively deployed over a 19-month time-frame in the North Sea, generating an unprecedented fine scale time-series of NPP in the region (Loveday and Smyth, 2020). At the large-scale, this time-series gives close agreement with existing satellite-based estimates of NPP for the region and previous in situ estimates. What has not been elucidated before is the high-frequency, small-scale, depth-resolved variability associated with bloom phenology, mesoscale phenomena and mixed layer dynamics.

scholarly journals Self-Supervised Pre-Training of Transformers for Satellite Image Time Series Classification

2020 ◽  
Author(s):  
Yuan Yuan ◽  
Lei Lin
Keyword(s):  
Time Series ◽  
Deep Learning ◽  
Large Scale ◽  
Temporal Structure ◽  
Satellite Image ◽  
Fine Tuning ◽  
Small Scale ◽  
Model Parameters ◽  
Learning Approaches ◽  
Wide Range

<div>Satellite image time series (SITS) classification is a major research topic in remote sensing and is relevant for a wide range of applications. Deep learning approaches have been commonly employed for SITS classification and have provided state-of-the-art performance. However, deep learning methods suffer from overfitting when labeled data is scarce. To address this problem, we propose a novel self-supervised pre-training scheme to initialize a Transformer-based network by utilizing large-scale unlabeled data. In detail, the model is asked to predict randomly contaminated observations given an entire time series of a pixel. The main idea of our proposal is to leverage the inherent temporal structure of satellite time series to learn general-purpose spectral-temporal representations related to land cover semantics. Once pre-training is completed, the pre-trained network can be further adapted to various SITS classification tasks by fine-tuning all the model parameters on small-scale task-related labeled data. In this way, the general knowledge and representations about SITS can be transferred to a label-scarce task, thereby improving the generalization performance of the model as well as reducing the risk of overfitting. Comprehensive experiments have been carried out on three benchmark datasets over large study areas. Experimental results demonstrate the effectiveness of the proposed method, leading to a classification accuracy increment up to 2.38% to 5.27%. The code and the pre-trained model will be available at https://github.com/linlei1214/SITS-BERT upon publication.</div><div><b>This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible.</b></div>

scholarly journals Complexity-Aware Simple Modeling

2018 ◽  
Author(s):  
Mariana Gómez-Schiavon ◽  
Hana El-Samad
Keyword(s):  
Mathematical Models ◽  
Large Scale ◽  
The Other ◽  
Small Scale ◽  
List Type ◽  
Unknown Parameters ◽  
New Approach ◽  
Complex Interactions ◽  
Scale Models ◽  
Simple Modeling

AbstractMathematical models continue to be essential for deepening our understanding of biology. On one extreme, simple or small-scale models help delineate general biological principles. However, the parsimony of detail in these models as well as their assumption of modularity and insulation make them inaccurate for describing quantitative features. On the other extreme, large-scale and detailed models can quantitatively recapitulate a phenotype of interest, but have to rely on many unknown parameters, making them often difficult to parse mechanistically and to use for extracting general principles. We discuss some examples of a new approach — complexity-aware simple modeling — that can bridge the gap between the small‐ and large-scale approaches.HighlightsSimple or small-scale models allow deduction of fundamental principles of biological systemsDetailed or large-scale models can be quantitatively accurate but difficult to analyzeComplexity-aware simple models can extract principles that are robust to the presence of unknown complex interactionsGraphical abstract

Sign in / Sign up

Export Citation Format

Share Document