An Approach of Time Series Piecewise Linear Representation Based on Local Maximum Minimum and Extremum

Time series is a kind of important complex data. It is a non-trivial problem to store data, analyze data and mine knowledge in its original data directly because of the inherent high dimensionality and complexity of the data. It is the most promising solution to achieve dimensionality reduction on the data. There are five major techniques in dimensionality reduction such as Discrete Fourier Transform, Discrete Wavelets Transform, Singular Value Decomposition, Symbolic Representation and Piecewise Linear Representation. Integrating the idea of important point and extreme point, a new approach of time series piecewise linear representation are proposed based on local maximum minimum and extremum in this paper. The results of experiments by using the public datasets from several different fields are shown that the proposed technique appears better fitting effect on the adjacent data value less volatile datasets and it has nice fitting effects in the volatile datasets under the low compression ratio condition compared with two other piecewise liner representation techniques.

[1]  Zhan Yan,et al.  Time Series Piecewise Linear Representation Based on Slope Extract Edge Point , 2006 .

[2]  Eamonn J. Keogh,et al.  Locally adaptive dimensionality reduction for indexing large time series databases , 2001, SIGMOD '01.

[3]  Eamonn J. Keogh,et al.  Dimensionality Reduction for Fast Similarity Search in Large Time Series Databases , 2001, Knowledge and Information Systems.

[4]  Eamonn J. Keogh,et al.  On the Need for Time Series Data Mining Benchmarks: A Survey and Empirical Demonstration , 2002, Data Mining and Knowledge Discovery.

[5]  Xiao Hui and Hi Yunfa.  Data Mining Based on Segmented Time Warping Distance in Time Series Database , 2005 .

[6]  Eugene Fink,et al.  Search for Patterns in Compressed Time Series , 2002, Int. J. Image Graph..

[7]  Christos Faloutsos,et al.  Efficiently supporting ad hoc queries in large datasets of time sequences , 1997, SIGMOD '97.

[8]  Davood Rafiei,et al.  On similarity-based queries for time series data , 1997, Proceedings 15th International Conference on Data Engineering (Cat. No.99CB36337).

[9]  Eamonn J. Keogh,et al.  A symbolic representation of time series, with implications for streaming algorithms , 2003, DMKD '03.

[10]  Christos Faloutsos,et al.  Efficient Similarity Search In Sequence Databases , 1993, FODO.

[11]  Haixun Wang,et al.  Landmarks: a new model for similarity-based pattern querying in time series databases , 2000, Proceedings of 16th International Conference on Data Engineering (Cat. No.00CB37073).

[12]  Christos Faloutsos,et al.  Fast Time Sequence Indexing for Arbitrary Lp Norms , 2000, VLDB.