algorithms for time series knowledge mining fabian moerchen 沈奕聰
TRANSCRIPT
![Page 1: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/1.jpg)
Algorithms For Time Series Knowledge Mining
Fabian Moerchen
沈奕聰
![Page 2: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/2.jpg)
Outline
• Introduction• Related work and motivation• Knowledge representation• Time series knowledge mining• Mining coincidence• Mining partial order
• Experiments• Discussion
![Page 3: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/3.jpg)
Introduction
• Backgroud• Patterns mined from symbolic interval data can provide
explanation for the underlying temporal processes or anomalous behavior• Symbolic interval time series are an important data format for
discovering temporal knowledge• Numerical time series are often converted to symbolic interval
time series
![Page 4: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/4.jpg)
Introduction
• Problems• Allen’s interval relations ‘s input usually consists of exact but
incomplete data and temporal constraints• Determining the consistency of the data• Answering queries about scenarios satisfying all constraints• Noisy and incorrect interval data
![Page 5: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/5.jpg)
Introduction
• Propose• Time Series Knowledge Representation(TSKR)• Hierarchical language• Based on interval time series• Extends the Unification-based Temporal Grammar• Using itemset techniques
![Page 6: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/6.jpg)
Related work and motivation
• Allen’s relations have severe disadvantages• Patterns from noisy interval data expressed with Allen’s interval relations are
not robust
![Page 7: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/7.jpg)
Related work and motivation
• Allen’s relations have severe disadvantages• Patterns expressed with Allen’s interval relations are ambiguous
![Page 8: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/8.jpg)
Related work and motivation
• Allen’s relations have severe disadvantages• Patterns expressed with Allen’s interval relations are not easily
comprehensible
![Page 9: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/9.jpg)
Related work and motivation
• The TSKR extends these core ideas achieving higher robustness and expressivity• The hierarchical structure of the UTG• The separation of temporal concepts
![Page 10: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/10.jpg)
Knowledge representationTones : basic primitives of the TSKR representing durationChord: a Chord pattern describes a time interval where k>0 Tones coincidePhrase: a paritial order of k>1 Chords
![Page 11: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/11.jpg)
Time series knowledge mining——Mining coincidence
![Page 12: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/12.jpg)
Time series knowledge mining——Mining coincidence
![Page 13: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/13.jpg)
Time series knowledge mining——Mining coincidence
![Page 14: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/14.jpg)
Time series knowledge mining——Mining partial order
![Page 15: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/15.jpg)
Time series knowledge mining——Mining partial order
![Page 16: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/16.jpg)
Experiments
![Page 17: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/17.jpg)
Experiments
![Page 18: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/18.jpg)
Experiments
![Page 19: Algorithms For Time Series Knowledge Mining Fabian Moerchen 沈奕聰](https://reader036.vdocuments.mx/reader036/viewer/2022062517/56649f335503460f94c507bf/html5/thumbnails/19.jpg)
Discussion
• Advantages• Hierarchical structure show the coinciding Tones and one Tone to show the
original numerical time series with the thresholds for discretization• The pruning by margin-closedness largely reduced the number of patterns • Effects on search space• Our novel data model conversion to itemset intervals greatly reduce the
redundancy• Search for phrases with our semantically motivated search space restrictions
are much faster than sequential pattern