Detection of Outliers in Time Series Data: A Frequency Domain Approach

2008 ◽  
Vol 1 (2) ◽  
pp. 130-137 ◽  
Author(s):  
O.I. Shittu ◽  
D.K. Shangodoyi
Keyword(s):  
Time Series ◽  
Frequency Domain ◽  
Time Series Data ◽  
Series Data ◽  
Frequency Domain Approach ◽  
Detection Of Outliers

scholarly journals Frequency–time analysis, low-rank reconstruction and denoising of turbulent flows using SPOD

2021 ◽  
Vol 926 ◽  
Author(s):  
Akhil Nekkanti ◽  
Oliver T. Schmidt
Keyword(s):  
Frequency Domain ◽  
Time Domain ◽  
Large Scale ◽  
Time Series Data ◽  
High Energy ◽  
Turbulent Jet ◽  
Low Rank ◽  
Series Data ◽  
Time Analysis ◽  
Frequency Domain Approach

Four different applications of spectral proper orthogonal decomposition (SPOD) are demonstrated on large-eddy simulation data of a turbulent jet. These are: low-rank reconstruction, denoising, frequency–time analysis and prewhitening. We demonstrate SPOD-based flow-field reconstruction using direct inversion of the SPOD algorithm (frequency-domain approach) and propose an alternative approach based on projection of the time series data onto the modes (time-domain approach). We further present a SPOD-based denoising strategy that is based on hard thresholding of the SPOD eigenvalues. The proposed strategy achieves significant noise reduction while facilitating drastic data compression. In contrast to standard methods of frequency–time analysis such as wavelet transform, a proposed SPOD-based approach yields a spectrogram that characterises the temporal evolution of spatially coherent flow structures. A convolution-based strategy is proposed to compute the time-continuous expansion coefficients. When applied to the turbulent jet data, SPOD-based frequency–time analysis reveals that the intermittent occurrence of large-scale coherent structures is directly associated with high-energy events. This work suggests that the time-domain approach is preferable for low-rank reconstruction of individual snapshots, and the frequency-domain approach for denoising and frequency–time analysis.

Characterising hydrodynamic controls on groundwater in a coastal urban aquifer using time and frequency domain responses at multiple spatiotemporal scales

2020 ◽  
Author(s):  
Ashley M Patton ◽  
Gabriel C Rau ◽  
Corinna Abesser ◽  
David R James ◽  
Peter J Cleall ◽  
...  
Keyword(s):  
Time Series ◽  
Frequency Domain ◽  
Time Series Data ◽  
Monitoring Network ◽  
Signal Propagation ◽  
Tidal River ◽  
Capital City ◽  
Series Data ◽  
Ocean Tide ◽  
Ocean Tides

<p>Urban environments often have highly variable and evolving hydrogeology. Coastal cities present even greater challenges to hydraulic and thermal conceptualisation and parameter estimation due to their complex dynamics and the heterogeneity of ocean-influenced hydraulic processes. Traditional methods of investigation (e.g. pump tests, invasive sampling) are time consuming, expensive, represent a snapshot in time and are difficult to conduct in built-up areas, yet properties derived from them are crucial for constructing models and forecasting urban groundwater evolution.</p><p>Here we present a novel approach to use passive sampling of groundwater head data to understand subsurface processes and derive hydraulic and geotechnical properties in an urban-coastal setting. This is illustrated using twenty years of high frequency (hourly) time-series data from an existing groundwater monitoring network comprising 234 boreholes distributed across Cardiff, the capital city of Wales, UK. We have applied Tidal Subsurface Analysis (TSA) to Earth, Atmospheric and Oceanic signals in groundwater time-series in the frequency domain, and also generated Barometric Response Functions in the time domain. By also observing the damping and attenuation of the response to ocean tides with distance from the coast and tidal rivers, this combination of analyses has enabled us to disentangle the influence of the different tidal components and estimate spatially distributed aquifer processes and parameters.</p><p>The data cover a period pre and post construction of a barrage across the coastline, impounding the city’s rivers. We were therefore able to observe a huge decrease in the subsurface ocean tide signal propagation following this human intervention, through the coastal and tidal river boundaries. These changes reveal variations in hydraulic responses and values of hydraulic diffusivity between different lithologies, notably with made-ground deposits being much less sensitive to ocean tides than the underlying sand and gravel aquifer. By being able to map the spatial variations in hydraulic response and barometric efficiency for the first time (and therefore formation compressibility and extent of aquifer confinement) we have been able to refine interpretations (and in some cases overcome misconceptions) derived from previous inferences made solely from borehole logs. We anticipate that linking the improved hydraulic characterisation, enabled by the new methodology, will also help better characterisation of the subsurface thermal regime, and management of shallow geothermal energy resources in coastal urban aquifers.</p>

scholarly journals A Review on Outlier/Anomaly Detection in Time Series Data

2021 ◽  
Vol 54 (3) ◽  
pp. 1-33
Author(s):  
Ane Blázquez-García ◽  
Angel Conde ◽  
Usue Mori ◽  
Jose A. Lozano
Keyword(s):  
Time Series ◽  
Outlier Detection ◽  
Time Series Data ◽  
State Of The Art ◽  
Series Data ◽  
Detection Techniques ◽  
The Past ◽  
Time Series Mining ◽  
Detection Of Outliers ◽  
Unsupervised Outlier Detection

Recent advances in technology have brought major breakthroughs in data collection, enabling a large amount of data to be gathered over time and thus generating time series. Mining this data has become an important task for researchers and practitioners in the past few years, including the detection of outliers or anomalies that may represent errors or events of interest. This review aims to provide a structured and comprehensive state-of-the-art on unsupervised outlier detection techniques in the context of time series. To this end, a taxonomy is presented based on the main aspects that characterize an outlier detection technique.

Method for obtaining response functions from noisy magnetotelluric data using frequency-domain independent component analysis

Geophysics ◽  
2021 ◽  
Vol 86 (1) ◽  
pp. E21-E35
Author(s):  
Shinya Sato ◽  
Tada-Nori Goto ◽  
Takafumi Kasaya ◽  
Hiroshi Ichihara
Keyword(s):  
Time Series ◽  
Independent Component Analysis ◽  
Frequency Domain ◽  
Time Series Data ◽  
Independent Component ◽  
Series Data ◽  
Response Functions ◽  
Cultural Noise ◽  
Subsurface Resistivity ◽  
Domain Independent

The magnetotelluric (MT) method has been used for visualizing subsurface resistivity structures and more recently for monitoring resistivity changes. However, electromagnetic data often include cultural noise, which can cause errors in the estimation of MT response functions and subsurface resistivity structure analysis. Frequency-domain independent component analysis (FDICA) offers advantages for MT data processing particularly because this method can extract hidden components in the observed data. These components can be decomposed into natural MT signals and cultural noise so that the noise effect in the recovered MT data is reduced. FDICA is applied to MT data acquired at the Kakioka Magnetic Observatory in Japan. The apparent resistivity and phase curves are obtained with small estimated errors between periods of 7 and 12,000 s, although the length of the time-series data is limited. The curves are smoother than those obtained using a conventional method. Various types of synthetic noise are added to the time series at Kakioka to test the noise-reduction performance of FDICA for MT data with high noise contamination. The results demonstrate that FDICA can be used to estimate MT response functions with high accuracy even under conditions in which more than half of the time-series data are contaminated by noise.

Detection of outliers in bivariate time series data

1987 ◽  
Vol 16 (12) ◽  
pp. 3701-3714 ◽  
Author(s):  
Ravindra Khattree ◽  
Dayanand N. Naik
Keyword(s):  
Time Series ◽  
Time Series Data ◽  
Series Data ◽  
Detection Of Outliers ◽  
Bivariate Time Series
Sign in / Sign up

Export Citation Format

Share Document