Local Correlation Tracking in Time Series

@article{Papadimitriou2006LocalCT,
  title={Local Correlation Tracking in Time Series},
  author={S. Papadimitriou and Jimeng Sun and Philip S. Yu},
  journal={Sixth International Conference on Data Mining (ICDM'06)},
  year={2006},
  pages={456-465}
}
We address the problem of capturing and tracking local correlations among time evolving time series. Our approach is based on comparing the local auto-covariance matrices (via their spectral decompositions) of each series and generalizes the notion of linear cross-correlation. In this way, it is possible to concisely capture a wide variety of local patterns or trends. Our method produces a general similarity score, which evolves over time, and accurately reflects the changing relationships… Expand
Frobenius correlation based u-shapelets discovery for time series clustering
TLDR
A dissimilarity score called FOTS whose computation is based on the eigenvector decomposition and the comparison of the autocorrelation matrices of the time series is proposed, which is robust to the presence of uncertainty. Expand
Low Redundancy Estimation of Correlation Matrices for Time Series Using Triangular Bounds
TLDR
This work introduces the novel problem of low redundancy estimation for correlation matrices to capture the complete correlation structure with as few parameters and correlation computations as possible and proposes a novel estimation algorithm that is very efficient and comes with formal approximation guarantees. Expand
Scale space multiresolution correlation analysis for time series data
TLDR
The results underline the usefulness of the scale space approach when the correlation between the time series exhibit time-varying structure in different scales. Expand
Finding Maximal Significant Linear Representation between Long Time Series
TLDR
A point-based approach is developed, which exploits a novel representation of linear correlation between time series on segments, and transforms the problem into geometric search, and presents a systematic empirical study to verify its efficiency and effectiveness. Expand
k-Shape: Efficient and Accurate Clustering of Time Series
The proliferation and ubiquity of temporal data across many disciplines has generated substantial interest in the analysis and mining of time series. Clustering is one of the most popular data miningExpand
Local correlation detection with linearity enhancement in streaming data
TLDR
A shape-based similarity measure is introduced into the framework, which refines the results by representative trend patterns to enhance the significance of linearity and improves the efficiency for online detection of correlation between data streams. Expand
Detecting Leaders from Correlated Time Series
TLDR
An efficient algorithm is proposed which is able to track the lagged correlation and compute the leaders incrementally, while still achieving good accuracy, and the detected leaders demonstrate high predictive power on the event of general time series entities, which can enlighten both climate monitoring and financial risk control. Expand
Robust online-surveillance of trend-coherence in multivariate data streams: the similar trend monitoring (STM) procedure
TLDR
The STM meets the high demands that come along with the online monitoring of multivariate data streams: it is fast to compute, robust against outliers, applicable when observations are missing, and does not require stationarity of the processes. Expand
Data Series Similarity Using Correlation-Aware Measures
TLDR
This study evaluates the plausibility and benefit of incorporating into the distance function measures of correlation, which enable us to capture the associations among neighboring values in the sequence, and proposes four such measures, inspired by statistical and probabilistic approaches, which can effectively model these correlations. Expand
Fast and Accurate Time-Series Clustering
TLDR
SBD, k-Shape, and k-MS emerge as domain-independent, highly accurate, and efficient methods for time-series comparison and clustering with broad applications, with combinations of the most competitive distance measures. Expand
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 46 REFERENCES
Optimal multi-scale patterns in time series streams
TLDR
This work introduces a method to discover optimal local patterns, which concisely describe the main trends in a time series and introduces a criterion to select the best window sizes, which most concisely capture the key oscillatory as well as aperiodic trends. Expand
Dimensionality Reduction and Forecasting on Streams
TLDR
The proposed method, SPIRIT, can incrementally find correlations and hidden variables, which summarise the key trends in the entire stream collection, which can be used to immediately spot potential anomalies, to do efficient forecasting and, more generally, to dramatically simplify further data processing. Expand
Fast Burst Correlation of Financial Data
TLDR
The proposed methods and data-structures can find applications for anomaly or novelty detection in telecommunications and network traffic, as well as in medical data. Expand
Knowledge Discovery from Heterogeneous Dynamic Systems using Change-Point Correlations
TLDR
This paper introduces a new nonlinear transformation, singular spectrum transformation (SST), to address the problem of knowledge discovery of causal relationships from a set of time series and demonstrates that SST enables us to discover a hidden and useful dependency between variables. Expand
Advanced spectral methods for climatic time series
TLDR
The connections between time series analysis and nonlinear dynamics, discuss signal-to-noise enhancement, and present some of the novel methods for spectral analysis are described. Expand
A multiresolution symbolic representation of time series
TLDR
This paper introduces a new representation of time series, the multiresolution vector quantized (MVQ) approximation, along with a new distance function that consistently outperforms all the major competitors achieving up to 20% better precision/recall and clustering accuracy on the tested datasets. Expand
A Novel Bit Level Time Series Representation with Implication of Similarity Search and Clustering
TLDR
This work introduces a new technique based on a bit level approximation of the data that allows raw data to be directly compared to the reduced representation, while still guaranteeing lower bounds to Euclidean distance. Expand
Indexing spatio-temporal trajectories with Chebyshev polynomials
TLDR
The Chebyshev polynomials are explored as a basis for approximating and indexing d-dimenstional trajectories and the key analytic result is the Lower Bounding Lemma, which shows that the Euclidean distance between two d-dimensional trajectories is lower bounded by the weighted Euclideans distance between the two vectors of ChebysHEv coefficients. Expand
StatStream: Statistical Monitoring of Thousands of Data Streams in Real Time
TLDR
This chapter solves the problem of maintaining multistream and time-delayed statistics in a scalable way that gives a guaranteed response time with high accuracy and Experiments conducted using synthetic and real data show that StatStream detects correlations efficiently and precisely. Expand
On Periodicity Detection and Structural Periodic Similarity
TLDR
Non-parametric methods for accurate periodicity detection and new periodic distance measures for time-series sequences are presented and introduced to assist in detecting, monitoring and visualizing structural periodic changes. Expand
...
1
2
3
4
5
...