Event Detection in Marine Time Series Data

Oehmcke, Stefan; Krämer, Oliver

doi:10.1007/978-3-319-24489-1_24

Cited by 20 publications

(12 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…LOF was initially designed to detect anomalies on spatial data [14]. But Oehmcke et al [66] extended the approach to use it also for time-series data.…”

Section: Local Outlier Factor (Lof)mentioning

confidence: 99%

Anomaly Detection in Univariate Time-series: A Survey on the State-of-the-Art

Braei,

Wagner

2020

Preprint

View full text Add to dashboard Cite

Anomaly detection for time-series data has been an important research field for a long time. Seminal work on anomaly detection methods has been focussing on statistical approaches. In recent years an increasing number of machine learning algorithms have been developed to detect anomalies on time-series. Subsequently, researchers tried to improve these techniques using (deep) neural networks. In the light of the increasing number of anomaly detection methods, the body of research lacks a broad comparative evaluation of statistical, machine learning and deep learning methods. This paper studies 20 univariate anomaly detection methods from the all three categories. The evaluation is conducted on publicly available datasets, which serve as benchmarks for time-series anomaly detection. By analyzing the accuracy of each method as well as the computation time of the algorithms, we provide a thorough insight about the performance of these anomaly detection approaches, alongside some general notion of which method is suited for a certain type of data.

show abstract

“…LOF was initially designed to detect anomalies on spatial data [14]. But Oehmcke et al [66] extended the approach to use it also for time-series data.…”

Section: Local Outlier Factor (Lof)mentioning

confidence: 99%

Anomaly Detection in Univariate Time-series: A Survey on the State-of-the-Art

Braei,

Wagner

2020

Preprint

View full text Add to dashboard Cite

show abstract

“…There are many other event detection methods developed in other application domains. Oehmcke et al [3] employed local outlier factor to detect events from marine time series data. To further improve results, dimensionality reduction methods are employed by the authors to the datasets.…”

Section: Related Workmentioning

confidence: 99%

“…There have been many existing machine learning methods to detect events out of time series datasets in various applications such as epileptic seizure detection from EEG signals and change detection from remotely sensed imagery datasets. Depending on the availability of labeled datasets, all these event detection methods for time series date sets can be categorized into supervised [3,4,5] and unsupervised methods [2,6,7]. Our study belongs to the supervised category, since we acquire labels for training and evaluation with the help of experts.…”

Section: Introductionmentioning

confidence: 99%

DeepDetect: A Cascaded Region-Based Densely Connected Network for Seismic Event Detection

Lin

Zhou

et al. 2019

IEEE Trans. Geosci. Remote Sensing

106

View full text Add to dashboard Cite

“…4) of the Institute for Chemistry and Biology of the Marine Environment in a tidal channel close to the island of Spiekeroog Badewien et al, 2009). The time-series station Spiekeroog ( The acquired data sets are fundamental for the improvement and validation of model results (Burchard and Badewien, 2015;Grashorn et al 2015;Lettman et al, 2009;Staneva et al, 2009;Burchard et al, 2008) as well as to answer various research questions (Rullkötter, 2009;Badewien et al, 2009;Hodapp et al, 2015;Meier et al, 2015;Holinde et al, 2015) and to improve fouling-prone sensing methods and quality assurance (Garaba et al, 2014;Schulz et al, 2015;Oehmcke et al, 2015).…”

Section: Pole Spiekeroogmentioning

confidence: 99%

The Coastal Observing System for Northern and Arctic Seas (COSYNA)

Baschek¹,

Schroeder²,

Brix³

et al. 2016

Preprint

View full text Add to dashboard Cite

Abstract. The Coastal Observing System for Northern and Arctic Seas (COSYNA) was established in order to better understand the complex interdisciplinary processes of northern seas and the arctic coasts in a changing environment. Particular focus is given to the German Bight in the North Sea as a prime example for a heavily used coastal area, and Svalbard as an example of an arctic coast that is under strong pressure due to global change. The automated observing and modelling system COSYNA is designed to monitor real time conditions, provide short-term forecasts and data products, and to assess the impact of anthropogenically induced change. Observations are carried out combining satellite and radar remote sensing with various in situ platforms. Novel sensors, instruments, and algorithms are developed to further improve the understanding of the interdisciplinary interactions between physics, biogeochemistry, and the ecology of coastal seas. New modelling and data assimilation techniques are used to integrate observations and models in a quasi-operational system providing descriptions and forecasts of key hydrographic variables. Data and data products are publically available free of charge and in real time. They are used by multiple interest groups in science, agencies, politics, industry, and the public.

show abstract

Event Detection in Marine Time Series Data

Cited by 20 publications

References 15 publications

Anomaly Detection in Univariate Time-series: A Survey on the State-of-the-Art

Anomaly Detection in Univariate Time-series: A Survey on the State-of-the-Art

DeepDetect: A Cascaded Region-Based Densely Connected Network for Seismic Event Detection

The Coastal Observing System for Northern and Arctic Seas (COSYNA)

Contact Info

Product

Resources

About