Detecting Outlying Time-Series with Global Alignment Kernels: Unterschied zwischen den Versionen

Aus IPD-Institutsseminar
Zur Navigation springen Zur Suche springen
Zeile 5: Zeile 5:
 
|betreuer=Florian Kalinke
 
|betreuer=Florian Kalinke
 
|termin=Institutsseminar/2020-12-11
 
|termin=Institutsseminar/2020-12-11
|kurzfassung=Using outlier detection algorithms, e.g., Support Vector Data Description (SVDD), for detecting outlying time-series usually requires extracting domain-specific attributes. However, this indirect way needs expert knowledge, making SVDD impractical for many real-world use cases. Incorporating “Global Alignment Kernels” directly into SVDD to compute the distance between time-series data bypasses the attribute-extraction step and makes the application of SVDD independent of the underlying domain.
+
|kurzfassung=Using outlier detection algorithms e.g., SVDD, for detecting outlying Time-Series usually requires extracting domain-specific attributes. However, this indirect way requires expert knowledge, which makes SVDD in many use cases impractical. Incorporating "Global Alignment Kernels" directly into SVDD to compute the distance between time-series data bypasses the attribute-extraction step and makes the application of SVDD independent of the underlying domain. In this work, a new algorithm based on "Global Alignment Kernels" and SVDD is developed. Its outlier detection capabilities will be evaluated on synthetic data as well as on real-world data sets. Additionally, our approach's performance will be compared to state-of-the-art methods for outlier detection, especially if our approach detects different outliers.
 
}}
 
}}

Version vom 2. Dezember 2020, 15:16 Uhr

Vortragende(r) Haiko Thiessen
Vortragstyp Proposal
Betreuer(in) Florian Kalinke
Termin Fr 11. Dezember 2020
Vortragsmodus
Kurzfassung Using outlier detection algorithms e.g., SVDD, for detecting outlying Time-Series usually requires extracting domain-specific attributes. However, this indirect way requires expert knowledge, which makes SVDD in many use cases impractical. Incorporating "Global Alignment Kernels" directly into SVDD to compute the distance between time-series data bypasses the attribute-extraction step and makes the application of SVDD independent of the underlying domain. In this work, a new algorithm based on "Global Alignment Kernels" and SVDD is developed. Its outlier detection capabilities will be evaluated on synthetic data as well as on real-world data sets. Additionally, our approach's performance will be compared to state-of-the-art methods for outlier detection, especially if our approach detects different outliers.