Downsampling Non-Uniformly Sampled Data

Eng, Frida; Gustafsson, Fredrik

doi:10.1155/2008/147407

Cited by 9 publications

(7 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For this data, the input is scalar x t k ∈ R, i.e., the input size m = 1, and the output d t k ∈ R, where d t k = x t k+1 . For the parameter selection, we perform a grid search on the number of hidden neurons and learning rate in the intervals q = [3,20] and η = [10 −3 , 10 −6 ], respectively. For the window size of the classical LSTM architecture with the preprocessing method, we search on the interval [( max /2), max ], where max equals to 10, 20, and 50 ms, respectively.…”

Section: A Regression Taskmentioning

confidence: 99%

“…In the classical data processing applications, data sequences are usually assumed to be uniformly sampled, and however, this is not the case in many real-life applications. For example, nonuniform sampling is used in many medical imaging applications [1], measurements in astronomy due to day and night conditions [2], and financial data [3], where the stock market values are redetermined by each transaction. Although nonuniformly sampled data frequently arises in these problems, there exist a few studies on nonuniformly sampled sequential data processing in neural networks [4], [5], machine learning [6], and signal processing literatures [7], [8].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Nonuniformly Sampled Data Processing Using LSTM Networks

Sahin

Kozat

2019

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

We investigate classification and regression for nonuniformly sampled variable length sequential data and introduce a novel long short-term memory (LSTM) architecture. In particular, we extend the classical LSTM network with additional time gates, which incorporate the time information as a nonlinear scaling factor on the conventional gates. We also provide forward-pass and backward-pass update equations for the proposed LSTM architecture. We show that our approach is superior to the classical LSTM architecture when there is correlation between time samples. In our experiments, we achieve significant performance gains with respect to the classical LSTM and phased-LSTM architectures. In this sense, the proposed LSTM architecture is highly appealing for the applications involving nonuniformly sampled sequential data.

show abstract

Section: A Regression Taskmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Nonuniformly Sampled Data Processing Using LSTM Networks

Sahin

Kozat

2019

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

show abstract

“…where ( ) are filter weights of our tested symmetric filters. The integrand interpolation method used the procedure from Eng and Gustafsson (2008), by which the irregular time series was filtered at regular positions ? (<@A) and the integrand of the filter-specific convolution integral ; (;<<) interpolated.…”

Section: Irregularitymentioning

confidence: 99%

Comparing methods for analysing time scale dependent correlations in irregularly sampled time series data

Reschke

Kunz

Laepple

2019

Computers & Geosciences

View full text Add to dashboard Cite

Time series derived from paleoclimate archives are often irregularly sampled in time and thus not analysable using standard statistical methods such as correlation analyses. Although measures for the similarity between time series have been proposed for irregular time series, they do not account for the time scale dependency of the relationship. Stochastically distributed temporal sampling irregularities act qualitatively as a low-pass filter reducing the influence of fast variations from frequencies higher than about 0.5 (∆ $%&) () , where ∆ $%& is the maximum time interval between observations. This may lead to overestimated correlations if the true correlation increases with time scale. Typically, correlations are underestimated due to a non-simultaneous sampling of time series. Here, we investigated different techniques to estimate time scale dependent correlations of weakly irregularly sampled time series, with a particular focus on different resampling methods and filters of varying complexity. The methods were tested on ensembles of synthetic time series that mimic the characteristics of Holocene marine sediment temperature proxy records. We found that a linear interpolation of the irregular time series onto a regular grid, followed by a simple Gaussian filter was the best approach to deal with the irregularity and account for the time scale dependence. This approach had both, minimal filter artefacts, particularly on short time scales, and a minimal loss of information due to filter length.

show abstract

“…However, nearly in every real life application, the data sequences usually contain missing input values due to various reasons such as inconvenience, anomalies and cost savings [2], [3]. Furthermore, in many real life problems such as medical imaging applications [4] and finance [5], we encounter nonuniformly sampled data, which can be modelled as a missing data case [6].…”

Section: Introduction a Preliminariesmentioning

confidence: 99%

A Tree Architecture of LSTM Networks for Sequential Regression with Missing Data

Sahin,

Kozat

2020

Preprint

View full text Add to dashboard Cite

We investigate regression for variable length sequential data containing missing samples and introduce a novel tree architecture based on the Long Short-Term Memory (LSTM) networks. In our architecture, we employ a variable number of LSTM networks, which use only the existing inputs in the sequence, in a tree-like architecture without any statistical assumptions or imputations on the missing data, unlike all the previous approaches. In particular, we incorporate the missingness information by selecting a subset of these LSTM networks based on "presence-pattern" of a certain number of previous inputs. From the mixture of experts perspective, we train different LSTM networks as our experts for various missingness patterns and then combine their outputs to generate the final prediction. We also provide the computational complexity analysis of the proposed architecture, which is in the same order of the complexity of the conventional LSTM architectures for the sequence length. Our method can be readily extended to similar structures such as GRUs, RNNs as remarked in the paper. In the experiments, we achieve significant performance improvements with respect to the state-of-the-art methods for the well-known financial and real life datasets.

show abstract

Downsampling Non-Uniformly Sampled Data

Cited by 9 publications

References 17 publications

Nonuniformly Sampled Data Processing Using LSTM Networks

Nonuniformly Sampled Data Processing Using LSTM Networks

Comparing methods for analysing time scale dependent correlations in irregularly sampled time series data

A Tree Architecture of LSTM Networks for Sequential Regression with Missing Data

Contact Info

Product

Resources

About