A New Time Series Similarity Measurement Method Based on the Morphological Pattern and Symbolic Aggregate Approximation

Yin, Jun; Wang, Rixin; Zheng, Huailiang; Yang, Yuantao; Li, Yuqing; Xu, Min

doi:10.1109/access.2019.2934109

Cited by 12 publications

(3 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Traditional distance metrics that are inspired by the concept of edit distance [10] include Lp-norms [65], Euclidean Distance [21], Dynamic Time Warping (DTW) [5], Longest Common Sub-sequence (LCSS) [58], Edit Sequence on Real Sequence (EDR) [11], Swale [38], Spatial Assembling Distance [13], etc. Further, several existing research papers have proposed different variants [16] of these traditional distance metrics for different objectives, such as run-time [17], applicability to specific problem [66], etc. Additionally, several recent research papers have proposed integration of both neighbourhood based metrics [24,27] and distance based metrics to train machine learning models, such as, SVM, Random Forest and ensemble models [34].…”

Section: Statistical Approachesmentioning

confidence: 99%

Finding Representative Sampling Subsets in Sensor Graphs using Time Series Similarities

Chakraborty¹,

Holm²,

Pedersen³

et al. 2022

Preprint

View full text Add to dashboard Cite

With the increasing use of IoT-enabled sensors, it is important to have effective methods for querying the sensors. For example, in a dense network of battery-driven temperature sensors, it is often possible to query (sample) just a subset of the sensors at any given time, since the values of the non-sampled sensors can be estimated from the sampled values. If we can divide the set of sensors into disjoint so-called representative sampling subsets that each represent the other sensors sufficiently well, we can alternate the sampling between the sampling subsets and thus, increase battery life significantly. In this paper, we formulate the problem of finding representative sampling subsets as a graph problem on a so-called sensor graph with the sensors as nodes. Our proposed solution, SubGraphSample, consists of two phases. In Phase-I, we create edges in the sensor graph based on the similarities between the time series of sensor values, analyzing six different techniques based on proven time series similarity metrics. In Phase-II, we propose two new techniques and extend four existing ones to find the maximal number of representative sampling subsets. Finally, we propose AutoSubGraphSample which auto-selects the best technique for Phase-I and Phase-II for a given dataset. Our extensive experimental evaluation shows that our approach can yield significant battery life improvements within realistic error bounds.

show abstract

Section: Statistical Approachesmentioning

confidence: 99%

Finding Representative Sampling Subsets in Sensor Graphs using Time Series Similarities

Chakraborty¹,

Holm²,

Pedersen³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…In order to obtain the information processing structure inside the monitoring system and extract the changing patterns of the time series, a symbolic processing method 37,38 is used to simplify the association rules of the time series. The symbolic processing method, which pays special emphasis on the deep understanding of the system rather than an accurate prediction, is a parallel branch with numerical sequence processing in the time series research field.…”

Section: Quantitative Indexes Of Association Rulesmentioning

confidence: 99%

Spatio‐temporal data mining method for joint cracks in concrete dam based on association rules

Huang

Liu

et al. 2021

Structural Contr & Hlth

View full text Add to dashboard Cite

Structural health monitoring (SHM) has been widely employed to reveal the hidden safety information and to diagnose the safety status in dam engineering fields. As one of the most important parameters of SHM, crack opening displacement (COD) is often used to evaluate the cracks or joints of concrete dams. In this paper, a new dam health analytic perspective is introduced by integrating the data mining method into SHM field, focusing on revealing the association rules in COD monitoring data. The association rules are investigated systematically, considering the cause-effect relations between external loads and structural response, the temporal characteristics of time series for a single sensor, the spatial characteristics of monitoring data for multisensors, and the abnormal characteristics for different items of structural responses. The association relation is quantified by proposing the quantitative indexes, including support degree, confidence degree, and promotion degree. The methods are used in the COD monitoring data of the Baishan concrete gravity-arch dam, which is located in a severely cold area in northeastern China. Results show that 4 out of 24 cause-effect association rules are extracted by calculating the association degree of monitored COD values, and 21 out of 24 crack sensors present a temporal association relationship, among which the confidence degree of two sensors reaches 100%. The variation trend of COD values is relevant with the locations of the crack sensors. These results are consistent with the dam safety monitoring theories and models, which would be very useful for extracting the SHM information between different sensors, predicting the trend of COD value and repairing the monitoring data series of COD sensors, or even for discovering an abnormal signal for the operation safety of dams.

show abstract

“…ED has been widely used because of its simple calculation and clear meaning [22]. However, it does not measure the trend of time series.…”

Section: B Time Series Similarity Measurementmentioning

confidence: 99%

A Novel Similarity Measurement and Clustering Framework for Time Series Based on Convolution Neural Networks

et al. 2020

View full text Add to dashboard Cite

In recent years, with the development of machine learning, especially after the rise of deep learning, time series clustering has been proven to effectively provide useful information in cloud computing and big data. However, many modern clustering algorithms are difficult to mine the complex features of time series, which is important for further analysis. Convolutional neural network provides powerful feature extraction capabilities and has excellent performance in classification tasks, but it is hard to be applied to clustering. Therefore, a similarity measurement method based on convolutional neural networks is proposed. This algorithm converts the number of output changes of the convolutional neural network in the same direction into the similarity of time series, so that the convolutional neural network can mine unlabeled data features in the clustering process. Especially by preferentially collecting a small amount of high similarity data to create labels, a classification algorithm based on the convolutional neural network can be used to assist clustering. The effectiveness of the proposed algorithm is proved by extensive experiments on the UCR time series datasets, and the experimental results show that its superior performance than other leading methods. Compared with other clustering algorithms based on deep networks, the proposed algorithm can output intermediate variables, and visually explain the principle of the algorithm. The application of financial stock linkage analysis provides an auxiliary mechanism for investment decision-making. INDEX TERMS time series, clustering, convolutional neural networks, similarity measurement

show abstract

A New Time Series Similarity Measurement Method Based on the Morphological Pattern and Symbolic Aggregate Approximation

Cited by 12 publications

References 31 publications

Finding Representative Sampling Subsets in Sensor Graphs using Time Series Similarities

Finding Representative Sampling Subsets in Sensor Graphs using Time Series Similarities

Spatio‐temporal data mining method for joint cracks in concrete dam based on association rules

A Novel Similarity Measurement and Clustering Framework for Time Series Based on Convolution Neural Networks

Contact Info

Product

Resources

About