Data Mining In Time Series Databases

Mark Last, Abraham Kandel, Horst Bunke9789812382900, 981-238-290-9

Adding the time dimension to real-world databases produces TimeSeries Databases (TSDB) and introduces new aspects and difficultiesto data mining and knowledge discovery. This book covers thestate-of-the-art methodology for mining time series databases. Thenovel data mining methods presented in the book include techniquesfor efficient segmentation, indexing, and classification of noisy anddynamic time series. A graph-based method for anomaly detection intime series is described and the book also studies the implicationsof a novel and potentially useful representation of time series asstrings. The problem of detecting changes in data mining models thatare induced from temporal databases is additionally discussed.

---

Table of contents :
Team-kb……Page 1
Contents……Page 12
Segmenting Time Series: A Survey And Novel Approach……Page 14
A Survey Of Recent Methods For Efficient Retrieval Of Similar Time Sequences……Page 36
Indexing Of Compressed Time Series……Page 56
Indexing Time-series Under Conditions Of Noise……Page 80
Change Detection In Classification Models Induced From Time Series Data……Page 114
Classification And Detection Of Abnormal Events In Time Series Of Graphs……Page 140
Þÿ……Page 162
Median Strings: A Review……Page 186