Browsing by Subject "Time series"

Now showing 1 - 9 of 9
  • Thumbnail Image
    Item
    Approximate search on massive spatiotemporal datasets.
    (2012-08) Brugere, Ivan
    Efficient time series similarity search is a fundamental operation for data exploration and analysis. While previous work has focused on indexing progressively larger datasets and has proposed data structures with efficient exact search algorithms, we motivate the need for approximate query methods that can be used in interactive exploration and as fast data analysis subroutines on large spatiotemporal datasets. This thesis formulates a simple approximate range query problem for time series data, and proposes a method that aims to quickly access a small number of high-quality results of the exact search resultset. We formulate an anytime framework, giving the user flexibility to return query results in arbitrary cost, where larger runtime incrementally improves search results. We propose an evaluation strategy on each query framework when the false dismissal class is very large relative to the query resultset and investigate the performance of indexing novel classes of time series subsequences.
  • Thumbnail Image
    Item
    Mapping oak wilt disease from space using land surface phenology
    (Remote Sensing of Environment, 2023-12-01) Guzmán, Jose A; Pinto-Ledezma, Jesús N; Frantz, David; Townsend, Philip A; Juzwik, Jennifer; Cavender-Bares, Jeannine
    Protecting the future of forests relies on our ability to observe changes in forest health. Thus, developing tools for sensing diseases in a timely fashion is critical for managing threats at broad scales. Oak wilt —a disease caused by a pathogenic fungus (Bretziella fagacearum)— is threatening oaks, killing thousands yearly while negatively impacting the ecosystem services they provide. Here we propose a novel workflow for mapping oak wilt by targeting temporal disease progression through symptoms using land surface phenology (LSP) from spaceborne observations. By doing so, we hypothesize that phenological changes in pigments and photosynthetic activity of trees affected by oak wilt can be tracked using LSP metrics derived from the Chlorophyll/Carotenoid Index (CCI). We used dense time-series observations from Sentinel-2 to create Analysis Ready Data across Minnesota and Wisconsin and to derive three LSP metrics: the value of CCI at the start and end of the growing season, and the coefficient of variation of the CCI during the growing season. We integrate high-resolution airborne imagery in multiple locations to select pixels (n = 3872) from the most common oak tree health conditions: healthy, symptomatic for oak wilt, and dead. These pixels were used to train an iterative Partial Least Square Discriminant (PLSD) model and derive the probability of an oak tree (i.e., pixel) in one of these conditions and the associated uncertainty. We assessed these models spatially and temporally on testing datasets revealing that it is feasible to discriminate among the three health conditions with overall accuracy between 80 and 82%. Within conditions, our models suggest that spatial variations among three CCI-derived LSP metrics can identify healthy (Area Under the Curve (AUC) = 0.98), symptomatic (AUC = 0.89), and dead (AUC = 0.94) oak trees with low false positive rates. The model performance was robust across different years as well. The predictive maps were used to guide local stakeholders to locate disease hotspots for ground verification and subsequent decision-making for treatment. Our results highlight the capabilities of LSP metrics from dense spaceborne observations to map diseases and to monitor large-scale change in biodiversity.
  • Thumbnail Image
    Item
    Mining dynamic relationships from spatio-temporal datasets: an application to brain fMRI data
    (2014-05) Atluri, Gowtham
    Spatio-temporal datasets are being widely collected in several domains such as climate science, neuorscience, sociology, and transportation. These data sets offer tremendous opportunities to address the imminent problems facing our society such as climate change, dementia, traffic congestion, crime etc. One example of a spatio-temporal dataset that is the focus of this dissertation is Functional Magnetic Resonance Imaging (fMRI) data. fMRI captures the activity at all locations in the brain and at regular time intervals. Using this data one can investigate the processes in the brain that relate to human psychological functions such as cognition, decision making etc. or physiological functions such as sensory perception or motor skills. Above all, one can advance the diagnosis and treatment procedures for mental disorders.The focus of this thesis is to study dynamic relationships between brain regions using fMRI data. Existing work in neuroscience has predominantly treated the relationships among brain regions as stationary. There is growing evidence in this community that the relationships between brain regions are transient. In the time series data mining community transient relationships have been studied and are shown to be useful for various tasks such as clustering and classification of time series data. In this work we focused on discovering combinations of brain regions that exhibit high similarity in the activity time series in small intervals. We proposed an efficient approach that can discover all such combinations exhaustively. We demonstrated its effectiveness on synthetic and real world data sets.We applied our approach on fMRI data collected in different settings on different groups of people and studied the reliability and replicability of the combinations we discover. Reliability is the degree to which a combination that is discovered using fMRI scans from a population can be found again using a different set of scans on the same population. Replicability is the degree to which a combination discovered using scans from one set of subjects can be discovered again using scans from a different set of subjects. These two factors reflect the generality of the combinations we discover. Our results suggest that the combinations we discover are indeed reliable and replicable. This indicates the validity of the combinations and they suggest that the underlying neuronal principles drive these combinations. We also investigated the utility of the combinations in studying differences between healthy and schizophrenia subjects.Existing work in estimating transient relationships among time series typically uses sliding time windows of a fixed length that are shifted from one end to the other using a fixed step size. This approach does not directly identify the intervals in which a pair of time series exhibit similarity. We proposed another computational approach to discover the time intervals where a given pair of time series are highly similar. We showed that our approach is efficient using synthetic datasets. We demonstrated the effectiveness of our approach on a synthetic dataset. Using this approach we provided a characterization of the transient nature of a relationship between time series and showed its utility in identifying task related transient connectivity in fMRI data that is collected while a subject is resting and while involved in a task.In summary, the computational approaches proposed in this thesis advance the state-of-the-art in time series data mining. Whereas the extensive evaluations that are performed on multiple fMRI datasets demonstrate the validity of the findings and provide novel hypothesis that can be systematically studied to advance the state-of-the-art in neuroscience.
  • Thumbnail Image
    Item
    Performance Evaluation of Different Detection Technologies for Signalized Intersections in Minnesota
    (Minnesota Department of Transportation, 2024-04) Grossman, Malcolm; Jiao, Yuankun; Hu, Haoji; Hourdos, John; Chiang, Yao-Yi
    This research evaluates the performance of non-intrusive detection technologies (NITs) for traffic signals in Minnesota. Prior work shows that while no single NIT device performs best in all situations, under specific circumstances, some NIT devices consistently outperform others. Our goal in this research is to find which NIT devices perform better in conditions specific to Minnesota and provide cost estimations and maintenance recommendations for operating these devices year-round. Our research has two main components: 1) synthesizing national and local experiences procuring, deploying, and maintaining NITs, and 2) evaluating real-world NIT deployments in Minnesota across different weather conditions. Our results and analysis combine the results from these steps to make recommendations informed by research and real-world experience operating NIT devices. Through interviews with Minnesota traffic signal operators, the research finds that environmental factors like wind, snow, and rain cause most NIT failures, requiring costly on-site maintenance. Operators emphasize the need for central monitoring systems, sun shields, and heated lenses to maintain performance. The research then analyzes NIT video, signal actuation, and weather data at six Twin Cities intersections using Iteris and Autoscope Vision technologies. No single NIT performs best, aligning with previous findings, but Autoscope Vision is less prone to lens blockages requiring on-site service. Our analysis also finds some intersections have more failures, indicating location and geometry impact performance. Key recommendations are based on the relative performance of a NIT in different weather conditions and accounting for local weather conditions when selecting a NIT at an intersection. We also recommend using central monitoring systems to troubleshoot remotely, installing heat shields to prevent snow/rain accumulation, and routine annual checks and checks after major storms.
  • Thumbnail Image
    Item
    Time series analysis of cardiometabolic parameters: reliability and energy drink response
    (2013-12) Nelson, Michael T.
    Cardiometabolic data is currently analyzed primarily by the use of averages. While this method can provide some data, further analysis by time series (variability) methods can provide more physiologic insights. Historically, time series analysis has been performed primarily using heart rate data in the form of heart rate variability (HRV) analysis. This was done to determine the status of the autonomic nervous system via changes in parasympathetic and sympathetic output. Researchers have used different methods of analysis, but a lack of reproducibility studies raises questions about the validity of these methods when applied to heart rate (HR) data. Currently in the literature, these methods have not applied to metabolic data such as the respiratory exchange ratio (RER). This dissertation will investigate the reliability of time series assessments of caridiometaoblic parameters. We hypothesize that in healthy individuals, HRV analysis performed on the same RR intervals but by two different measurement systems, are indeed interchangeable. We further hypothesize that the time series analysis of metabolic data such as the RER will be stable and repeatable over two trials conducted under the same conditions. Lastly, we hypothesize that under conditions of physical stress (e.g. ride time-to-exhaustion) and biochemical stress (e.g. energy drink), resting HR and HR variability preexercise will be altered and the ride time-to-exhaustion will be increased after subjects consume an energy drink (standardized to 2.0mg/kg caffeine) compared to a taste-matched placebo. The results of this dissertation will provide further insight into the repeatability of these time series analyses, which could be utilized for future research to determine metabolic flexibility.
  • Thumbnail Image
    Item
    Time series analysis of cardiometabolic parameters: reliability and energy drink response
    (2013-12) Nelson, Michael Thomas
    Cardiometabolic data is currently analyzed primarily by the use of averages. While this method can provide some data, further analysis by time series (variability) methods can provide more physiologic insights. Historically, time series analysis has been performed primarily using heart rate data in the form of heart rate variability (HRV) analysis. This was done to determine the status of the autonomic nervous system via changes in parasympathetic and sympathetic output. Researchers have used different methods of analysis, but a lack of reproducibility studies raises questions about the validity of these methods when applied to heart rate (HR) data. Currently in the literature, these methods have not applied to metabolic data such as the respiratory exchange ratio (RER). This dissertation will investigate the reliability of time series assessments of caridiometaoblic parameters. We hypothesize that in healthy individuals, HRV analysis performed on the same RR intervals but by two different measurement systems, are indeed interchangeable. We further hypothesize that the time series analysis of metabolic data such as the RER will be stable and repeatable over two trials conducted under the same conditions. Lastly, we hypothesize that under conditions of physical stress (e.g. ride time-to-exhaustion) and biochemical stress (e.g. energy drink), resting HR and HR variability preexercise will be altered and the ride time-to-exhaustion will be increased after subjects consume an energy drink (standardized to 2.0mg/kg caffeine) compared to a taste-matched placebo. The results of this dissertation will provide further insight into the repeatability of these time series analyses, which could be utilized for future research to determine metabolic flexibility.
  • Thumbnail Image
    Item
    Time series change detection: algorithms for land cover change.
    (2010-04) Boriah, Shyam
    The climate and earth sciences have recently undergone a rapid transformation from a data-poor to a data-rich environment. In particular, climate and ecosystem related observations from remote sensors on satellites, as well as outputs of climate or earth system models from large-scale computational platforms, provide terabytes of temporal, spatial and spatio-temporal data. These massive and information-rich datasets offer huge potential for advancing the science of land cover change, climate change and anthropogenic impacts. One important area where remote sensing data can play a key role is in the study of land cover change. Specifically, the conversion of natural land cover into human-dominated cover types continues to be a change of global proportions with many unknown environmental consequences. In addition, being able to assess the carbon risk of changes in forest cover is of critical importance for both economic and scientific reasons. In fact, changes in forests account for as much as 20% of the greenhouse gas emissions in the atmosphere, an amount second only to fossil fuel emissions. Thus, there is a need in the earth science domain to systematically study land cover change in order to understand its impact on local climate, radiation balance, biogeochemistry, hydrology, and the diversity and abundance of terrestrial species. Land cover conversions include tree harvests in forested regions, urbanization, and agricultural intensification in former woodland and natural grassland areas. These types of conversions also have significant public policy implications due to issues such as water supply management and atmospheric CO2 output. In spite of the importance of this problem and the considerable advances made over the last few years in high-resolution satellite data, data mining, and online mapping tools and services, end users still lack practical tools to help them manage and transform this data into actionable knowledge of changes in forest ecosystems that can be used for decision making and policy planning purposes. In particular, previous change detection studies have primarily relied on examining differences between two or more satellite images acquired on different dates. Thus, a technological solution that detects global land cover change using high temporal resolution time series data will represent a paradigm-shift in the field of land cover change studies. To realize these ambitious goals, a number of computational challenges in spatio-temporal data mining need to be addressed. Specifically, analysis and discovery approaches need to be cognizant of climate and ecosystem data characteristics such as seasonality, non-stationarity/inter-region variability, multi-scale nature, spatio-temporal autocorrelation, high-dimensionality and massive data size. This dissertation, a step in that direction, translates earth science challenges to computer science problems, and provides computational solutions to address these problems. In particular, three key technical capabilities are developed: (1) Algorithms for time series change detection that are effective and can scale up to handle the large size of earth science data; (2) Change detection algorithms that can handle large numbers of missing and noisy values present in satellite data sets; and (3) Spatio-temporal analysis techniques to identify the scale and scope of disturbance events.
  • Thumbnail Image
    Item
    Time series segmentation techniques for land cover change detection
    (2013-05) Garg, Ashish
    Ecosystem-related observations from remote sensors on satellites offer a significant possibility for understanding the location and extent of global land cover change. In this study, we focus on time series segmentation techniques in the context of land cover change detection. We propose a model based time series segmentation algorithm inspired by an event detection framework proposed in the field of statistics. We also present a novel model free change detection algorithm for detecting land cover change that is computationally simple, efficient, non-parametric and takes into account the inherent variability present in the remote sensing data. A key advantage of this method is that it can be applied globally for a variety of vegetation without having to identify the right model for specific vegetation types. We evaluate the change detection capacity of the proposed techniques on both synthetic and MODIS EVI data sets. We illustrate the importance and relative ability of different algorithms to account for the natural variation in the EVI data set.
  • Thumbnail Image
    Item
    Topics on Climate Model Output Analyses
    (2021-10) Gong, Kaibo
    Comparison of two different data samples, and of paired data samples, is a well known problem in Statistics. Specifically, there is a wide range of applications in the fields of climate study. In this thesis, we provide a brief review on the ensemble of climate models and the need of probabilistic evaluation of model outputs, which is equivalent to the comparison between two models. Based on recent advancements in the context of evaluating climate model outputs, we develop two different approaches for comparing two functional time series. The first one is based on wavelet decomposition and the second one by comparing the local spectral density of non-stationary series. For the last chapter, we conduct a brief review on Gaussian Process and a framework for Bayesian Optimization, which establishes a theoretical framework and algorithmic properties of t-process based spatio-temporal modeling, for further use in modeling climate and neuroscience data.