Search NASA⌕ Search

SEARCH · Search NASA

Results for “Multivariate time series”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 73 records · Page 4

Spatial estimation from remotely sensed data via empirical Bayes models

Multichannel satellite image data, available as LANDSAT imagery, are recorded as a multivariate time series (four channels, multiple passovers) in two spatial dimensions. The application of parametric empirical Bayes theory to classification of, and estimating the probability of, each crop type at each of a large number of pixels is considered. This theory involves both the probability distribution of imagery data, conditional on crop types, and the prior spatial distribution of crop types. For the latter Markov models indexed by estimable parameters are used. A broad outline of the general theory reveals several questions for further research. Some detailed results are given for the special case of two crop types when only a line transect is analyzed. Finally, the estimation of an underlying continuous process on the lattice is discussed which would be applicable to such quantities as crop yield.

Hill, J. R.↗

Biogeochemical Response to Mesoscale Physical Forcing in the California Current System

In the first part of the project, we investigated the local response of the coastal ocean ecosystems (changes in chlorophyll, concentration and chlorophyll, fluorescence quantum yield) to physical forcing by developing and deploying Autonomous Drifting Ocean Stations (ADOS) within several mesoscale features along the U.S. west coast. Also, we compared the temporal and spatial variability registered by sensors mounted in the drifters to that registered by the sensors mounted in the satellites in order to assess the scales of variability that are not resolved by the ocean color satellite. The second part of the project used the existing WOCE SVP Surface Lagrangian drifters to track individual water parcels through time. The individual drifter tracks were used to generate multivariate time series by interpolating/extracting the biological and physical data fields retrieved by remote sensors (ocean color, SST, wind speed and direction, wind stress curl, and sea level topography). The individual time series of the physical data (AVHRR, TOPEX, NCEP) were analyzed against the ocean color (SeaWiFS) time-series to determine the time scale of biological response to the physical forcing. The results from this part of the research is being used to compare the decorrelation scales of chlorophyll from a Lagrangian and Eulerian framework. The results from both parts of this research augmented the necessary time series data needed to investigate the interactions between the ocean mesoscale features, wind, and the biogeochemical processes. Using the historical Lagrangian data sets, we have completed a comparison of the decorrelation scales in both the Eulerian and Lagrangian reference frame for the SeaWiFS data set. We are continuing to investigate how these results might be used in objective mapping efforts.

Niiler, Pearn P.↗

Temporal Variability of Observed and Simulated Hyperspectral Earth Reflectance

The Climate Absolute Radiance and Refractivity Observatory (CLARREO) is a climate observation system designed to study Earth's climate variability with unprecedented absolute radiometric accuracy and SI traceability. Observation System Simulation Experiments (OSSEs) were developed using GCM output and MODTRAN to simulate CLARREO reflectance measurements during the 21st century as a design tool for the CLARREO hyperspectral shortwave imager. With OSSE simulations of hyperspectral reflectance, Feldman et al. [2011a,b] found that shortwave reflectance is able to detect changes in climate variables during the 21st century and improve time-to-detection compared to broadband measurements. The OSSE has been a powerful tool in the design of the CLARREO imager and for understanding the effect of climate change on the spectral variability of reflectance, but it is important to evaluate how well the OSSE simulates the Earth's present-day spectral variability. For this evaluation we have used hyperspectral reflectance measurements from the Scanning Imaging Absorption Spectrometer for Atmospheric Cartography (SCIAMACHY), a shortwave spectrometer that was operational between March 2002 and April 2012. To study the spectral variability of SCIAMACHY-measured and OSSE-simulated reflectance, we used principal component analysis (PCA), a spectral decomposition technique that identifies dominant modes of variability in a multivariate data set. Using quantitative comparisons of the OSSE and SCIAMACHY PCs, we have quantified how well the OSSE captures the spectral variability of Earth?s climate system at the beginning of the 21st century relative to SCIAMACHY measurements. These results showed that the OSSE and SCIAMACHY data sets share over 99% of their total variance in 2004. Using the PCs and the temporally distributed reflectance spectra projected onto the PCs (PC scores), we can study the temporal variability of the observed and simulated reflectance spectra. Multivariate time series analysis of the PC scores using techniques such as Singular Spectrum Analysis (SSA) and Multichannel SSA will provide information about the temporal variability of the dominant variables. Quantitative comparison techniques can evaluate how well the OSSE reproduces the temporal variability observed by SCIAMACHY spectral reflectance measurements during the first decade of the 21st century. PCA of OSSE-simulated reflectance can also be used to study how the dominant spectral variables change on centennial scales for forced and unforced climate change scenarios. To have confidence in OSSE predictions of the spectral variability of hyperspectral reflectance, it is first necessary for us to evaluate the degree to which the OSSE simulations are able to reproduce the Earth?s present-day spectral variability.

Roberts, Yolanda↗

A Novel Machine Learning Method for Surface PM2.5 Estimations from Geostationary Satellites

Particulate matter (PM) with a diameter of less or equal to 2.5 μm, known as PM , affects human health as it penetrates the respiratory system. The Environmental Protection Agency (EPA) measures the atmospheric concentration of PM using air quality monitors stationed throughout the Continental United States (CONUS). Such measurements are points on a spatial domain and therefore, might not be representative of the air quality at nearby areas considering that the composition of the atmosphere is highly variable from place to place. Satellite based AOD permits a spatially uniform means of estimating PM and new geostationary satellites provide high temporal and spatial resolution estimation of AOD. However, the concentration of PM is non-linearly dependent on other atmospheric parameters that include relative humidity, temperature, and height of the planetary boundary layer. This information may be estimated at similar spatial and temporal resolutions as AOD from numerical modeling such as from the National Oceanic and Atmospheric Administration’s (NOAA) High Resolution Rapid Refresh (HRRR) model which resolves near real-time atmospheric conditions over the CONUS. The estimation of PM concentration is a multi-parametric problem that considers the effect of temporal dependencies among the different parameters. Deep learning approaches are appropriate for such complex estimation problems as they intrinsically capture relations among multiple non-linear parameters. This study compares deep-learning methods to traditional regression analysis to demonstrate the capabilities of these methods in predicting PM2.5 concentrations. Additionally, a novel ensemble learning approach is employed to identify scientific processes that could further improve the estimation of PM concentration. Utilizing Long Short-Term Memory (LSTM) neural networks, which are suitable for multivariate time series estimation problems as they are capable of learning long-term dependencies, individual models are created for each EPA station and trained on the aforementioned dataset collocated over each station. Individual station models are merged if the model's performance is improved by reducing the root mean squared error (RMSE) metric. This ensemble training method ultimately reduces the RMSE value. Evaluation of these results provide insights into physical processes and related observable parameters that may contribute to PM concentrations. Identified parameters evaluated to be statistically different between the merged and unmerged models are expected to improve overall performance. These new parameters are then utilized for reevaluation of the deep learning methods with an extreme gradient boosting model with an RMSE of 5.5 providing the best results.

George Priftis↗

Anomaly Detection in Flight Operational Data Using Deep Learning

In this session, we demonstrate two recently developed deep learning models for anomaly detection in flight operational data by the Data Sciences Group at NASA Ames Research Center. The first model is Convolutional Variational Auto-Encoder (CVAE) [1], which is an unsupervised deep encoder-decoder model, designed specifically for finding anomalies in heterogeneous multivariate time series data. We will demonstrate its application to finding anomalies in streaming data from NASA’s Digital Information Platform’s Fuser source. CVAE identifies data instances that are not representative of expected nominal behavior as anomalous. Since it is an unsupervised approach, the flagged anomalies will need to be reviewed by the subject matter experts (SMEs) for validation and labeling and is designed to assist with vulnerability discovery within Safety Monitoring System programs. The second model is Robust and Explainable Semi-supervised Anomaly Detection (RESAD) model [2], which builds on CVAE to allow learning from both minimally labeled data (previously reviewed by the SMEs) as well as majority unlabeled data. RESAD takes advantage of graph theoretic techniques to propagate the labels from the labeled data to the unlabeled data based on a pre-defined similarity metric and structures the learned feature space from flight time-series so that data of the same class would cluster tightly together. This model characteristic is enabled by training with an augmented loss function and allows learning of a more informative feature space for down-stream tasks such as search and active learning. We demonstrate RESAD using data from the NASA DASHlink project [3].

anomaly detection↗

Reliable machine prognostic health management in the presence of missing data

Prognostics and health management enables the prediction of future degradation and remaining useful life (RUL) for in-service systems based on historical and contemporary data, showing promise for many practical applications. One major challenge for prognostics is the common occurrence of missing values in time-series data, often caused by disruptions in sensor communication or hardware/software failures. Another major concern is that the sufficient prior knowledge of critical component degradation with a clear failure threshold is often not readily available in practice. These issues can significantly hinder the application of advanced signal and data analysis methods and consequently degrade the health management performance. In this article, we propose a novel data-driven framework that is capable of providing accurate and reliable predictions of degradation and RUL. In this approach, one-hot health state indicators are appended to the historical time series so that the model learns end-of-life automatically. A modified gate recurrent unit based variational autoencoder is employed in generative adversarial networks to model the temporal irregularity of the incomplete time series. Furthermore, experiments on multivariate time-series datasets collected from real-world aeroengines verify that significant performance improvement can be achieved using the proposed model for robust long-term prognostics.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗

Visualizing Clustering and Uncertainty Analysis with Multivariate Longitudinal Data

Multivariate time-series datasets are intrinsic to the study of dynamic, naturalistic behavior, such as in the applications of finance and motion video analysis. Statistical models provide the ability to identify event patterns in these data under conditions of uncertainty, but researchers must be able to evaluate how well a model uses available information in a dataset for clustering decisions and for uncertainty information. The Hidden Markov Model (HMM) is an established method for clustering time-series data, where the hidden states of the HMM are the clusters. We develop novel methods for quantifying the uncertainty of the performance of and for visualizing the clustering performance and uncertainty of fitting a HMM to multivariate time-series data. We explain the usefulness of uncertainty quantification and visualization with evaluating the performance of clustering models, as well as how information exploitation of time-series datasets can be enhanced. We implement our methods to cluster patterns of scanpaths from raw eye tracking data.

97 MATHEMATICS AND COMPUTING↗

Online evolutionary neural architecture search for multivariate non-stationary time series forecasting

Time series forecasting (TSF) is one of the most important tasks in data science. TSF models are usually pre-trained with historical data and then applied on future unseen datapoints. However, real-world time series data is usually non-stationary and models trained offline usually face problems from data drift. Models trained and designed in an offline fashion can not quickly adapt to changes quickly or be deployed in real-time. To address these issues, this work presents the Online NeuroEvolution-based Neural Architecture Search (ONE-NAS) algorithm, which is a novel neural architecture search method capable of automatically designing and dynamically training recurrent neural networks (RNNs) for online forecasting tasks. Without any pre-training, ONE-NAS utilizes populations of RNNs that are continuously updated with new network structures and weights in response to new multivariate input data. ONE-NAS is tested on real-world, large-scale multivariate wind turbine data as well as the univariate Dow Jones Industrial Average (DJIA) dataset. These results demonstrate that ONE-NAS outperforms traditional statistical time series forecasting methods, including online linear regression, fixed long short-term memory (LSTM) and gated recurrent unit (GRU) models trained online, as well as state-of-the-art, online ARIMA strategies. Additionally, results show that utilizing multiple populations of RNNs which are periodically repopulated provide significant performance improvements, allowing this online neural network architecture design and training to be successful.

97 MATHEMATICS AND COMPUTING↗

qFeature

This package contains statistical routines for extracting features from multivariate time-series data which can then be used for subsequent multivariate statistical analysis to identify patterns and anomalous behavior. It calculates local linear or quadratic regression model fits to moving windows for each series and then summarizes the model coefficients across user-defined time intervals for each series. These methods are domain agnostic-but they have been successfully applied to a variety of domains, including commercial aviation and electric power grid data.

Source record↗

qFeature

This package contains statistical routines for extracting features from multivariate time-series data which can then be used for subsequent multivariate statistical analysis to identify patterns and anomalous behavior. It calculates local linear or quadratic regression model fits to moving windows for each series and then summarizes the model coefficients across user-defined time intervals for each series. These methods are domain agnostic-but they have been successfully applied to a variety of domains, including commercial aviation and electric power grid data.

Source record↗

Latent Neural ODE for Integrating Multi-Timescale Measurements in Smart Distribution Grids

Under a smart grid paradigm, there has been an increase in sensor installations to enhance situational awareness. The measurements from these sensors can be leveraged for real-time monitoring, control, and protection. However, these measurements are typically irregularly sampled. These measure-ments may also be intermittent due to communication bandwidth limitations. To tackle this problem, this paper proposes a novel latent neural ordinary differential equations (LODE) approach to aggregate the unevenly sampled multivariate time-series measurements. The proposed approach is flexible in performing both imputations and predictions while being computationally efficient. Simulation results on IEEE 37 bus test systems illustrate the efficiency of the proposed approach.

multi time-scale measurements↗

Missing-Data Nonparametric Coherency Estimation

Chave recently proposed an estimator for multitaper spectral density where the time series contains missing values. In this article we generalize this technique to a multitaper estimator of coherence and phase and show that one can also obtain bootstrapped confidence intervals. Additionally, we give two examples. The first is a toy example in which the true coherence is known. In the second example we show that the multitaper missing-data coherence estimator computed on real data with a single gap comprising 11% of the data outperforms the Daniell-smoothed coherence estimator where there are no gaps. The case where the two time series have different missing indices is also discussed.

42 ENGINEERING↗

Fast Multivariate Search on Large Aviation Datasets

Multivariate Time-Series (MTS) are ubiquitous, and are generated in areas as disparate as sensor recordings in aerospace systems, music and video streams, medical monitoring, and financial systems. Domain experts are often interested in searching for interesting multivariate patterns from these MTS databases which can contain up to several gigabytes of data. Surprisingly, research on MTS search is very limited. Most existing work only supports queries with the same length of data, or queries on a fixed set of variables. In this paper, we propose an efficient and flexible subsequence search framework for massive MTS databases, that, for the first time, enables querying on any subset of variables with arbitrary time delays between them. We propose two provably correct algorithms to solve this problem (1) an R-tree Based Search (RBS) which uses Minimum Bounding Rectangles (MBR) to organize the subsequences, and (2) a List Based Search (LBS) algorithm which uses sorted lists for indexing. We demonstrate the performance of these algorithms using two large MTS databases from the aviation domain, each containing several millions of observations Both these tests show that our algorithms have very high prune rates (>95%) thus needing actual

Bhaduri, Kanishka↗

Detecting outliers in streaming time series data from ARM distributed sensors

The Atmospheric Radiation Measurement (ARM) Data Center at ORNL collects data from a number of permanent and mobile facilities around the globe. The data is then ingested to create high level scientific products. High frequency streaming measurements from sensors and radar instruments at ARM sites require high degree of accuracy to enable rigorous study of atmospheric processes. Outliers in collected data are common due to instrument failure or extreme weather events. Thus, it is critical to identify and flag them. We employed multiple univariate, multivariate and time series techniques for outlier detection methods and studied their effectiveness. First, we examined Pearson correlation coefficient which is used to measure the pairwise correlations between variables. Singular Spectrum Analysis (SSA) was applied to detect outliers by removing the anticipated annual and seasonal cycles from the signal to accentuate anomalies. K-means was applied for multivariate examination of data from collection of sensors to identify any deviation from expected and known patterns and identify abnormal observation. The Pearson correlation coefficient, SSA and K-means methods were later combined together in a framework to detect outliers through a range of checks. We applied the developed method to data from meteorological sensors at ARM Southern Great Plains site and validated against existing database of known data quality issues.

Lu, Yuping↗

GenAI-Based Digital Twins Aided Data Augmentation Increases Accuracy in Real-Time Cokurtosis-Based Anomaly Detection of Wearable Data

Early detection of potential infectious disease outbreaks is crucial for developing effective interventions. In this study, we introduce advanced anomaly detection methods tailored for health datasets collected from wearables, offering insights at both individual and population levels. Leveraging real-world physiological data from wearables, including heart rate and activity, we developed a framework for the early detection of infection in individuals. Despite the availability of data from recent pandemics, substantial gaps remain in data collection, hindering method development. To bridge this gap, we utilized Wasserstein Generative Adversarial Networks (WGANs) to generate realistic synthetic wearable data, augmenting our dataset for training. Subsequently, we use these augmented datasets to implement a cokurtosis-based technique for anomaly detection in multivariate time-series data. Our approach includes a comprehensive assessment of uncertainties in synthetic data compared to the actual data upon which it was modeled, as well as the uncertainty associated with fine-tuning anomaly detection thresholds in physiological measurements. Through our work, we present an enhanced method for early anomaly detection in multivariate datasets, with promising applications in healthcare and beyond. This framework could revolutionize early detection strategies and significantly impact public health response efforts in future pandemics.

Data-Driven Digital Twins↗

Parallel hybrid quantum-classical machine learning for kernelized time-series classification

Supervised time-series classification garners widespread interest because of its applicability throughout a broad application domain including finance, astronomy, biosensors, and many others. Here, in this work, we tackle this problem with hybrid quantum-classical machine learning, deducing pairwise temporal relationships between time-series instances using a timeseries Hamiltonian kernel (TSHK). A TSHK is constructed with a sum of inner products generated by quantum states evolved using a parameterized time evolution operator. This sum is then optimally weighted using techniques derived from multiple kernel learning. Because we treat the kernel weighting step as a differentiable convex optimization problem, our method can be regarded as an end-to-end learnable hybrid quantum-classical-convex neural network, or QCC-net, whose output is a data set-generalized kernel function suitable for use in any kernelized machine learning technique such as the support vector machine (SVM). Using our TSHK as input to a SVM, we classify univariate and multivariate time-series using quantum circuit simulators and demonstrate the efficient parallel deployment of the algorithm to 127-qubit superconducting quantum processors using quantum multi-programming.

97 MATHEMATICS AND COMPUTING↗

A hidden Markov model for population-level cervical cancer screening data

The Cancer Registry of Norway has been administrating a national cervical cancer screening program since 1992 by coordinating triennial cytology exam screenings for the female population between 25 and 69 years of age. Up to 80% of cancers are prevented through mass screening, but this comes at the expense of considerable screening activity and leads to overtreatment of clinically asymptomatic precancers. Here, we present a continuous-time, time-inhomogeneous hidden Markov model which was developed to understand the screening process and cervical cancer carcinogenesis in detail. By leveraging 1.7 million individual's multivariate time-series of medical exams performed over a 25-year period, we simultaneously estimate all model parameters. We show that an age-dependent model reflects the Norwegian screening program by comparing empirical survival curves from observed registry data and data simulated from the proposed model. The model can be generalized to include more detailed individual-level covariates as well as new types of screening exams. By utilizing individual screening histories and covariate data, the proposed model shows potential for improving strategies for cancer screening programs by personalizing recommended screening intervals.

59 BASIC BIOLOGICAL SCIENCES↗